[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16196783#comment-16196783 ] Ishan Chattopadhyaya commented on SOLR-10317: - I'm back from a vacation and now have some time to work on this. This is still not ready and needs about a week of dev effort (full time). I'm starting to work on this part time and without any help I should be able to do this in two weeks, i.e. by 21-22 October. Btw, Vivek and I spoke about the benchmarking suite last month: https://youtu.be/44v2WljG1R0?t=19m59s > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, SOLR-10317.patch, > SOLR-10317.patch, Screenshot from 2017-07-30 20-30-05.png, > changes-lucene-20160907.json, changes-solr-20160907.json, managed-schema, > solrconfig.xml > > > Currently hosted at: http://212.47.242.214/MergedViewCloud.html > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16159975#comment-16159975 ] Ishan Chattopadhyaya commented on SOLR-10317: - [~viveknarang], thank you for your work on the project. I hope you can continue to contribute to this effort even after GSoC. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16159974#comment-16159974 ] Ishan Chattopadhyaya commented on SOLR-10317: - An update on the project status here: The GSoC is over, and the project is not complete yet. Since the last month of the three month program was unproductive, the project has officially failed. However, I am working on the missing parts. The major piece is the configurability; every benchmark is currently hardcoded and extending/modifying them is extremely difficult. I shall try to have something for review in a few days. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16146931#comment-16146931 ] Vivek Narang commented on SOLR-10317: - Hello [~ichattopadhyaya], JSON based configurable querying is working now. Yes, I can provide a written summary of these updates. Regards. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16146591#comment-16146591 ] Ishan Chattopadhyaya commented on SOLR-10317: - Sounds great. bq. I would like to give you a quick demo soon. Regards. Can you write up something and/or record a screencast? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16146458#comment-16146458 ] Vivek Narang commented on SOLR-10317: - Hello [~ichattopadhyaya], Now the configurable approach to indexing is working. Querying will also start working in the same way in a couple of hours from now. I would like to give you a quick demo soon. Regards. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16145941#comment-16145941 ] ASF GitHub Bot commented on SOLR-10317: --- GitHub user viveknarang opened a pull request: https://github.com/apache/lucene-solr/pull/240 SOLR-10317 SOLR-10317: Integrate new benchmark suite [https://issues.apache.org/jira/browse/SOLR-10317] You can merge this pull request into a Git repository by running: $ git pull https://github.com/viveknarang/lucene-solr SolrNightlyBenchmarks Alternatively you can review and apply these changes as the patch at: https://github.com/apache/lucene-solr/pull/240.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #240 commit 2f6665b9ef8d218ffdd1116da7f65df1b183541d Author: Vivek NarangDate: 2017-06-08T02:31:26Z SolrNightlyBenchmarks font change ... commit 1713a476f8c044ed47349bb22e32a673213272b4 Author: Vivek Narang Date: 2017-06-08T06:30:10Z lastrun folder naming bug fix ... commit 33e1b1839919e6f8ce367462636a18ed3d5aaee7 Author: Vivek Narang Date: 2017-06-08T07:12:47Z duplicate commit in queue processing bug fix ... commit eccf60ca2af17352b521cce6cd1b5e7400906d84 Author: Vivek Narang Date: 2017-06-09T00:47:29Z Adding liscense header ... commit 3aa6c6b4b1dac1fdf228a5c2caab6df3c3b36a48 Author: Vivek Narang Date: 2017-06-10T14:22:16Z modifying Util class commit ce3165f18d40a38e11534f46bd98ca97116b591b Author: Vivek Narang Date: 2017-06-11T03:24:43Z AND-OR numeric queries commit d1f1ac75938c9d240cc21d38af26c7315ad1d65c Author: Vivek Narang Date: 2017-06-11T07:53:50Z CRITICAL CPU HEAP Grap Overlap bugs fixed batch ... commit 0c91cc85c93e4f56a30595283a73db400e46a762 Author: Vivek Narang Date: 2017-06-11T10:59:32Z Tab text change + First Commit: Qtime Percentile ... commit 13c674b9523d508742702046dfbc25798f9c8064 Author: Vivek Narang Date: 2017-06-11T19:59:28Z Formatting view for environment data for each point on popup ... commit e90357be4696d55d3a9060dca924cc62148747d5 Author: Vivek Narang Date: 2017-06-11T20:18:26Z Changing the legend on the pages ... commit 721d913ae33594779a2f0b680d5556f5c81e7895 Author: Vivek Narang Date: 2017-06-12T00:28:24Z MAJOR: Cloud Indexing various configurations (212, 221, 313, 422) implementation ... commit 2d38444e948e8b926edd06ee347719388ddf7e1e Author: Vivek Narang Date: 2017-06-12T00:39:16Z Navigation divider update ... commit 628780c918fd9e85251d1e4b955e09fb403dd698 Author: Vivek Narang Date: 2017-06-12T01:07:29Z Webapp file bug fix ... commit 406fa718504c26224b383237e3d67c78aabb25ab Author: Vivek Narang Date: 2017-06-12T01:28:25Z updating test plans uncommenting tests ... commit 796d118ed896b078a7b5bf26602f43e29ab69ac6 Author: Vivek Narang Date: 2017-06-12T21:49:02Z Folder creation bug fix. Some formatting ... commit fc2ff1d329d862fdcbbf3890397db28458646dfa Author: Vivek Narang Date: 2017-06-13T02:45:54Z AND OR numeric query bug fix ... commit 494555fa69a0592678544a077c77d7c0236a60de Author: Vivek Narang Date: 2017-06-14T05:02:11Z Update README.md commit 5c148c8ad8903da0735a9995e3f56dcd84c2179f Author: Vivek Narang Date: 2017-06-14T05:07:12Z Update commit 696e74cc4d021a7c870e1c01a751f0b4a8d03e9d Author: Vivek Narang Date: 2017-06-14T05:08:41Z Create README.md commit e57ebed359b3b0589539dfa4bfe29ee34d411616 Author: Vivek Narang Date: 2017-06-14T05:11:42Z Create README.md commit 63eaa45ae9ff575914a0bfc52950c8747c511980 Author: Vivek Narang Date: 2017-06-14T05:12:29Z Create README.md commit 79aaee63ca5ac5e5fcb5aab7fecd58509e0d587a Author: Vivek Narang Date: 2017-06-14T05:18:43Z Create README.md commit 506b6ae4481d5313b78ab47471bb20a50f535e41 Author: Vivek Narang Date: 2017-06-14T05:20:52Z Create README.md commit 9e98cb2cc936905afefc67553bf1e0c07f4e99da Author: Vivek Narang Date: 2017-06-14T05:25:10Z Create README.md commit ce82a96727603b6c40b5f4b57366de25282a34c9 Author: Vivek Narang Date: 2017-06-14T05:25:35Z Create README.md
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16145707#comment-16145707 ] Ishan Chattopadhyaya commented on SOLR-10317: - Vivek, unless I'm mistaken, the deadline of the GSoC has now passed. Can you please make a final submission? Either raise a Github PR or a attach a patch here. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16145631#comment-16145631 ] Ishan Chattopadhyaya commented on SOLR-10317: - Hi [~viveknarang], Do you have any status, please? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16141250#comment-16141250 ] Vivek Narang commented on SOLR-10317: - Refactoring is being done section by section, for now, two parts - QueryClient and Metric have been modified. Will add the required documentation soon. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16141212#comment-16141212 ] Ishan Chattopadhyaya commented on SOLR-10317: - I see the branch just has pseudo code based on the approach (to make the suite configurable) we discussed last Sunday over the pair programming session (which is as per the comment here [0]). What's the status for actual working code? Also, can you please document here the refactoring to frontend (to make this suite configurable) that we discussed on that call? [0] - https://issues.apache.org/jira/browse/SOLR-10317?focusedCommentId=16115774=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-16115774 > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16137488#comment-16137488 ] Vivek Narang commented on SOLR-10317: - Hello [~ichattopadhyaya], please switch to "SolrNightlyBenchmarks-R2" branch for the latest updates. Regards. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16135892#comment-16135892 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya], I was working on a big section of backend refactoring last week, I will send an update after committing the code tonight. Regards. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16132135#comment-16132135 ] Ishan Chattopadhyaya commented on SOLR-10317: - Vivek, are there any updates, please? Full week as passed without any progress or updates. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16125810#comment-16125810 ] Vivek Narang commented on SOLR-10317: - 10-4 Captain! > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16125790#comment-16125790 ] Ishan Chattopadhyaya commented on SOLR-10317: - Ah, 12h5m from now? :-) I'll stay up for the meeting. Ping me if you're available any time before that. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16125786#comment-16125786 ] Ishan Chattopadhyaya commented on SOLR-10317: - Sure, I'm ready. :-) > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16125739#comment-16125739 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya], Let's have a Google hangout meeting today (8/14) at 11 pm EST. Thanks > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16124776#comment-16124776 ] Vivek Narang commented on SOLR-10317: - Hey [~ichattopadhyaya], I think an awesome dynamic UI mechanism idea just struck me. i'll summarize it soon. Regards > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16124514#comment-16124514 ] Ishan Chattopadhyaya commented on SOLR-10317: - Thanks for the update, [~viveknarang]. Would you available for a call to discuss the plan/approaches on Monday (or any other time) over a voice call/hangout? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16124510#comment-16124510 ] Vivek Narang commented on SOLR-10317: - The front-end plan is not clear yet. I will update as soon as I have more clarity on the approach. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16124494#comment-16124494 ] Ishan Chattopadhyaya commented on SOLR-10317: - What is the plan for the frontend redesign, so as to make it work with your above backend redesign? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16124490#comment-16124490 ] Vivek Narang commented on SOLR-10317: - These CSV files are needed at the front-end side. 1. Primary Consumer of CSV files - The dygraph API. 2. Format changes with the required graph layout. [Col1 - X-axis, Col2,Col3 ... etc are the Y-axis]. The first line in the CSV files is the visible label as seen on the graphs. 3. Dygraph uses these CSV data files. 4. These CSV files are referenced to the calling JS functions from the dygraph API, and then they are handled by the dygraph API internally. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16124482#comment-16124482 ] Ishan Chattopadhyaya commented on SOLR-10317: - Can you please elaborate: # why these CSV files are needed, # what is the format, # who uses it # how it is used? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16124417#comment-16124417 ] Vivek Narang commented on SOLR-10317: - Hello [~ichattopadhyaya], Initial steps for backend refactoring (up to the point of creating CSV files): Top layer - JSON meta data file, followed by creating a list of BenchmarkConfiguration objects based on parsed information. A BenchmarkConfiguration object holds the information on a specific benchmark which is then passed across several layers inside providing information at each level. Against each benchmark, there is a map object that holds the results that can then be passed to the point that creates the CSV files. This approach will remove a major portion of partially redundant code in the Tests.java as well as remove the enums that were limiting configuration at various levels. Please check sample code on a new branch that tries to showcase a part of this approach. [https://github.com/viveknarang/lucene-solr/commit/0869a7c2aa131de6afdca38e04fa4627f87d6a46] Regards. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16123952#comment-16123952 ] Ishan Chattopadhyaya commented on SOLR-10317: - > If you need help in making the frontend dynamic, just give me a shout out. > I'll get you some expert to guide you. [~upayavira], in case [~viveknarang] needs some help with the UI of this tool, would it be possible for you to help him a bit (in advisory/mentoring capacity)? I am a bit challenged in frontend design and development, and I'm afraid that I'll be of very little help to him if he needs it. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16123940#comment-16123940 ] Ishan Chattopadhyaya commented on SOLR-10317: - Any update, please? [~viveknarang] > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121960#comment-16121960 ] Ishan Chattopadhyaya commented on SOLR-10317: - Discuss the design here before you start implementing. We don't want you to design another inflexible system. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121941#comment-16121941 ] Ishan Chattopadhyaya commented on SOLR-10317: - Good to know you are better. Just do it. You have 19 days left in the project. https://developers.google.com/open-source/gsoc/timeline If you need help in making the frontend dynamic, just give me a shout out. I'll get you some expert to guide you. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16121865#comment-16121865 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya], I am feeling better now, Thanks! I took this time to think about your suggestions and here is my opinion. Yes, it is a good idea to make this configurable but there are some challenges to it. Even if the backend (leading to the point where CSVs are generated) is made dynamic (which is not as straight forward as it seems), the front end will have to be made dynamic which violates the initial agreement of keeping the front end static as discussed initially. Limitations on the front-end are one of the reasons why the implementation is the way it is. We need to make the front end dynamic for this idea to be realistic. The suite even in the current state is extensible but for that one will have to modify the classes and the front-end to add/modify benchmarks. However, one can always find better ways to do things and JSON based configurable suite is a good idea as an improvement! I want this benchmark project to be useful for the community and I would not recommend waiting for another year to make the suggested changes. However, since I anticipate these changes to be time taking and since the deadline is just over a week away, I recommend creating a duplicate branch of the current project and implement your suggestions over the coming weeks. Regards > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16118880#comment-16118880 ] Ishan Chattopadhyaya commented on SOLR-10317: - Thanks for informing, Vivek. bq. I have been unwell for last three days Hope you get well soon, since the suite is coming along really nicely so far. However, when you are unable to work, kindly inform as soon as possible. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16118865#comment-16118865 ] Vivek Narang commented on SOLR-10317: - Hello [~ichattopadhyaya], I have been unwell for last three days, will try to resume work as soon as possible. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16118796#comment-16118796 ] David Smiley commented on SOLR-10317: - bq. Unbxd Inc., the company I am doing Solr consulting for currently, has generously decided to support this effort by bearing the cost of bare metal servers that Vivek is currently using. Awesome and thanks! I'm looking forward to seeing periodic benchmarks we can all access. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16118794#comment-16118794 ] Ishan Chattopadhyaya commented on SOLR-10317: - [~viveknarang], you have not posted the daily status for Monday, 7 August. Have you reviewed the above concerns and my comments about making the benchmarking suite more configurable? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16118791#comment-16118791 ] Ishan Chattopadhyaya commented on SOLR-10317: - Unbxd Inc., the company I am doing Solr consulting for currently, has generously decided to support this effort by bearing the cost of bare metal servers that Vivek is currently using. These include the two C2L Scaleway servers (https://www.scaleway.com/pricing) that Vivek has been using for past two months, and shall be continuing to use for on-going development. Apart from that, Unbxd will also cover the cost of hardware to publicly host the nightly benchmarking service, on stable bare metal servers (Scaleway, Packet.net etc.), going forward. The specifics of the latter, i.e. the type of boxes we need for this, coordination required with Apache's infra team etc. are TBD. For a background, Unbxd uses Apache Solr and hosts a large number of collections (for e-commerce search and discovery) and is excited to support the continuous benchmarking of SolrCloud (along with the various plugins developed in-house). > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16115774#comment-16115774 ] Ishan Chattopadhyaya commented on SOLR-10317: - https://github.com/viveknarang/lucene-solr/blob/SolrNightlyBenchmarks/dev-tools/solrnightlybenchmarks/src/main/java/org/apache/solr/tests/nightlybenchmarks/TestPlans.java#L32-L34 {code} public enum BenchmarkTestType { PROD_TEST, DEV_TEST } {code} There should be no concept of "prod" or "dev". It should be a benchmark that relies on configuration rather than assumed defaults like "prod" or "dev". Also, I don't think we should be using terms like "tests" for individual benchmarks. There is no verification or assertions in these benchmarks, other than just timing data collection. The way the entire code is laid out, it is extremely hard to add new benchmarks. It might require an entirely new GSoC project next year to make this useful for the community. Hard coded test scenarios, really? https://github.com/viveknarang/lucene-solr/blob/SolrNightlyBenchmarks/dev-tools/solrnightlybenchmarks/src/main/java/org/apache/solr/tests/nightlybenchmarks/MetricCollector.java#L31-L219 This mess absolutely needs to be fixed at the earliest. I my opinion, the way the benchmarks in this suite should configured as: {code} { "index-benchmarks": [ { "name": "CLOUD_INDEXING", "description": "some shit", "replication-type": "cloud", "dataset-file": "filename containing data" "setups": [ { "collection": "cloud_2x2", "replicationFactor": 2, "shards": 2, "min-threads": 1, "max-threads": 16 }, { "collection": "cloud_1x1", "replicationFactor": 1, "shards": 1, "min-threads": 1, "max-threads": 16 }, { "collection": "cloud_1x2", "replicationFactor": 2, "shards": 1, "min-threads": 1, "max-threads": 16 } ] }, { "name": "CLOUD_PARTIAL_UPDATE", "description": "some shit", "replication-type": "cloud", "dataset-file": "filename containing full documents", "updates-file": "filename containing updates", "setups": [ { "collection": "partial_2x2", "replicationFactor": 2, "shards": 2, "min-threads": 1, "max-threads": 16 }, { "collection": "partial_1x1", "replicationFactor": 1, "shards": 1, "min-threads": 1, "max-threads": 16 }, { "collection": "partial_1x2", "replicationFactor": 2, "shards": 1, "min-threads": 1, "max-threads": 16 } ] }, ... more such benchmarks ... ], "query-benchmarks": [ { "name": "TERM_NUMERIC_QUERY_CLOUD_2T", "description": "some shit describing the benchmark", "replication-type": "cloud or standalone""collection/core": "", "query-file": "name of file containing all the queries for this benchmark", "client-type": "CUSC or CSC or HSC etc.", "min-threads": 1"max-threads": 8 }, ... more such benchmarks ... ] } {code} Based on this, the suite will do the right thing. Various things to consider here: # Partial updates benchmarks should: {code} for every replicationFactor, shards, thread combination: Create a new collection with given name, and given replicationFactor and shards Index the full dataset without timing. Start timer Update all documents Stop timer, record difference in time Delete this collection. {code} # Full document indexing benchmarks should: {code} for every replicationFactor, shards, thread combination: Create a new collection with given name, and given replicationFactor and shards Start timer Index the documents Stop timer, record difference in time if (numThread != maxThread): Delete this collection else // don't delete, since this collection needs to stay for query benchmarks {code} # For every query benchmark: {code} for every collection, thread combination: Stop and start all Solr nodes (so that caches are cleared) Wait till all replicas for the given collection are "active" Issue around 100-200 queries to warm up the searchers. Start timer Query the collection using the given numThreads Stop timer, record difference in time {code} In case there's any information that the graphs need, but not covered here, please comment/discuss. What do you think of the above proposal (in general or in specific parts) to make the suite easier to configure/manage/extend? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project:
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16115754#comment-16115754 ] Ishan Chattopadhyaya commented on SOLR-10317: - This is just ridiculous: {code} if (nodes == 2 && shards == "1" && replicas == "2") { ... } else if (nodes == 2 && shards == "2" && replicas == "1") { ... } else if (nodes == 3 && shards == "1" && replicas == "3") { ... } else if (nodes == 4 && shards == "2" && replicas == "2") { {code} https://github.com/viveknarang/lucene-solr/blob/SolrNightlyBenchmarks/dev-tools/solrnightlybenchmarks/src/main/java/org/apache/solr/tests/nightlybenchmarks/Tests.java#L301 All of these should be configurable and easily extendible (extensible?). > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16115749#comment-16115749 ] Ishan Chattopadhyaya commented on SOLR-10317: - A huge problem with the benchmarking suite is that it *takes 6.5 hours* to run one benchmarking run, end to end, *irrespective of how fast the machine is*! This clearly shows there's some major problem with how these benchmarks are implemented. I'd expect the benchmarks to finish quicker on a faster machine. For the benchmarks that the suite currently runs, I find it very odd that it takes so long to run, end to end. Can you please explain why this happens? I'm planning to take a deeper look into the code soon to figure out why. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16115745#comment-16115745 ] Ishan Chattopadhyaya commented on SOLR-10317: - *Kindly*, be specific in your status reports. Phrases like "new QPS calculation logic" or "bugs fixed" have no meaning unless you link to a Github issue describing your "new .. logic" or the exact bugs you fixed. Ideally, those issues should be linked to your exact commits for fixing/implementing them. Also, please fix the following parameter names: --commit-id could be as simple as --commit, --test-with-number-of-documents could be as simple as --num-docs, --use-sample-dataset could be as simple as --sample Also, what is the difference between --use-sample-dataset and --test-with-number-of-documents? Does the latter only control the index size? This description, {code} * --use-sample-dataset X.XX Use this option when you want to work in dev-mode (i.e while enhancing /debugging this project.). Please also pass a value in the range 0.01 to 1 with this parameter. This value is the percentage of data set that is used in this mode. {code} could be: {code} * --use-sample-dataset X.XX This value (0 - 1.0) is the fraction of dataset & queries used for indexing and querying. Useful in dev-mode. Default value, when this parameter is not specified, is 1.0 {code} > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16115261#comment-16115261 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya], Status: -Testing complete and bugs fixed in the new QPS calculation logic, the code will be pushed in one hour. Regards. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16113867#comment-16113867 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya], Status for Today: - Testing the new QPS calculation logic. Regards > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16112115#comment-16112115 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya], Status for Today: Testing and refactoring QPS calculation logic - controlling the number of queries instead of the time period. Regards > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16110222#comment-16110222 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya] Status For Today: Conducted tests and added more logging & Null checks. After doing the RCA on the issue observed recently, I have come to a conclusion that a recent commit done in Solr, is the cause for this issue. Please see this: [https://github.com/apache/lucene-solr/commit/67b3d4e108b82b56ca45b9e78bcac52bd2280c8b]. This is the commit after which the issue started appearing (the ones you noted a few comments before). If you observe lines 128 and 149 (_default/conf/managed-schema), you see that *_pi is removed and now *_i must be used in instead of it. Work in progress: - Adding an option for dev-tests where a small data set and small query set will allow for quicker execution of the benchmark framework, speeding up the development/debugging process. - Adding more checks in the code. Regards > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16108330#comment-16108330 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya] Status for today: Testing, adding more logging and RCA for the observed errors. I see that the test is running fine for an older commit eg: 5f1c5ce81286ab2e0c0f98c47d332511ea57e8c1 while there are some errors noted when running for a recent commit eg: 3a405971b9e06e2004e0d66ae1b82f530de969f2. Regards > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106760#comment-16106760 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya], Work status as of now is as follows: - Remove Internal queuing capability: Done - Rename Root and webapp directory: Done - Create a script for processing commits for last n days: In progress - Missing exception chaining/handling: Done, pending verification. - Log4j implementation: Done. (solrnightlybenchmarks.log file created in /logs folder) - Null pointer exception handling across all class files: In progress - Providing shell script for download: Done (download.sh script in /scripts folder) - Embedded Jetty server support, removing external HTTP server dependency: Done As of now, I am currently testing against the changes made today. Regards > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106602#comment-16106602 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya] I have added issues on the list. Thanks for pointing out important gaps. I am working on them and will provide updates soon. Regards > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106566#comment-16106566 ] Ishan Chattopadhyaya commented on SOLR-10317: - Had a discussion with Vivek over Hangout, and we discussed all these issues. Vivek will create Github issues (on his repository) to track all these issues, and fix all of them asap. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > Screenshot from 2017-07-30 20-30-05.png, SOLR-10317.patch, SOLR-10317.patch, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106525#comment-16106525 ] Ishan Chattopadhyaya commented on SOLR-10317: - bq. Such exceptions should be fatal and the entire benchmarking suite should crash out after such exceptions. Silently ignoring them seems like a horrible thing to do. https://github.com/viveknarang/lucene-solr/blob/SolrNightlyBenchmarks/dev-tools/SolrNightBenchmarks/src/main/java/org/apache/solr/tests/nightlybenchmarks/QueryClient.java#L477-L479 The individual benchmarks should crash on failures. Failures indicate some problem with either Solr or the benchmarking suite. Failures due to the latter should be discovered asap. Failures due to the latter must not exist, and lets fix all such possibilities. In the above case, you're polling from a queue and not even checking if the poll was successful; hence the NPE. Please fix, and please stop ignoring exceptions! > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106507#comment-16106507 ] Ishan Chattopadhyaya commented on SOLR-10317: - My suggestion, as I have discussed with you before, would be to use Log4J (or any other similar logging framework). You can easily log INFO, DEBUG, ERROR and WARNING level messages. Also, the logs would be preserved even after the runs finish. It seems that currently, the only logging that is happening is at the console and hence the logs are not preserved after the run finishes. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106499#comment-16106499 ] Ishan Chattopadhyaya commented on SOLR-10317: - How can I see logs for these runs? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106384#comment-16106384 ] Ishan Chattopadhyaya commented on SOLR-10317: - {quote} >...which property in that properties file needs to be adjusted? How to > enable/disable this ability etc.? More detailed steps coming up shortly. {quote} Just remove that ability. The benchmarking suite need not download files on its own. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106383#comment-16106383 ] Ishan Chattopadhyaya commented on SOLR-10317: - {quote} >That link is useless. How will the user download all the files from that > link? Link Removed. {quote} I think you are not understanding what I'm trying to convey above. I need you to provide *clear instructions* on *how to download the data manually*. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106376#comment-16106376 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya] bq. Perhaps "Benchmarking data files" or "Data files". Will Rename. bq. That link is useless. How will the user download all the files from that link? Link Removed. bq. Downloading files requires a specialized software, and the benchmarking suite should not be concerned with that. Resuming partial downloads, multi-threaded downloading, etc. are special features that the downloader should take care of. I think the business of downloading files should be left for the user to deal with. Shell script coming up shortly. bq. ...which property in that properties file needs to be adjusted? How to enable/disable this ability etc.? More detailed steps coming up shortly. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106364#comment-16106364 ] Ishan Chattopadhyaya commented on SOLR-10317: - {quote} >You need to be specific in terms of steps the user needs to take. https://github.com/viveknarang/lucene-solr/tree/SolrNightlyBenchmarks/dev-tools/SolrNightBenchmarks#steps-to-launch The steps have always existed on this file! Do you find any steps missing? {quote} What existed and exists now is "The system has the ability to check and download required data files from the source, as specified in the properties file". Again, why would a user care about the system's ability to download files? What he needs to know is, how to invoke that ability. Which "properties file", which property in that properties file needs to be adjusted? How to enable/disable this ability etc.? {quote} I don't think there is any complexity here. Checking the existence of these files is already an essential step. The files are only downloaded when missing and essentially a wget is done to download files please see code below. If you still insist I will create a shell script for this. {quote} Checking for existence is good. But automatically downloading it is unnecessary complexity. "essentially a wget is done to download files" <--- why the f*** would you want to invoke a shell command from within Java, when you can have the user/admin do it before launching your service? Downloading files requires a specialized software, and the benchmarking suite should not be concerned with that. Resuming partial downloads, multi-threaded downloading, etc. are special features that the downloader should take care of. I think the business of downloading files should be left for the user to deal with. {quote} >Secondly, "please use the link provided below" is equivalent to telling > the user to "go f*** yourself" The link has been provided as an option and is not a required step. When the user has configured the system to automatically download required files, there is no need for the user to manually download files. {quote} That link is useless. How will the user download all the files from that link? {quote} > Why are you calling it "test data files"? Please suggest what you want these files to be referred as. {quote} Perhaps "Benchmarking data files" or "Data files". > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106352#comment-16106352 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya] bq. Your README file [0] is unreadable due to horizontal scrollers in each section. Fixed now please check. [https://github.com/viveknarang/lucene-solr/tree/SolrNightlyBenchmarks/dev-tools/SolrNightBenchmarks] > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106295#comment-16106295 ] Vivek Narang commented on SOLR-10317: - [~ichattopadhyaya] I have sent a detailed status email for this week combined with last week a few minutes before. Since major current and future planned goals are already completed by now (as also noted in the email) the TODO items as of now are: Code refactoring, working on some of the points noted by you above and thorough testing of the platform. I will provide additional updates shortly. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106240#comment-16106240 ] Ishan Chattopadhyaya commented on SOLR-10317: - I'd like to add, however, that *overall, the project looks to be in great shape*, and you have done a good job at it so far. The second month evaluation has just completed, and you have passed that evaluation. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106238#comment-16106238 ] Ishan Chattopadhyaya commented on SOLR-10317: - So far, the project has relied on some manual steps to deploy a webapp on a webserver. This is unnecessary complexity on the part of the admin who is setting it up. Can you please use a Jetty based embedded web-server, which someone can run using an ant or mvn command? Upon starting this embedded web-server, it can serve all the UI and reports through a pre-configured port. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106234#comment-16106234 ] Ishan Chattopadhyaya commented on SOLR-10317: - bq. I have fixed the label bug on graphs. Please see http://212.47.242.214/MergedViewStandalone.html Thanks bq. I will compile and list the TODO items shortly. What is the status of this? As per GSoC guidelines, a student is expected to put in about 40 hours of work each week (which is approximately 8 hours a day). Couldn't you spare 5 minutes from those 8-16 hours (in last 2 days) to get to this? I requested you to share the weekly status with the mentors via email (which you have been doing, even though the level of detail can be improved). Going forward, can you kindly post the status here and include sufficient details on work you have done each week? bq. The ability to locate and download test data files is awesome and now the system will be able to check the data file presence and download the missing file(s) by itself as required. The source URL is configurable through the properties file. This is just unnecessary complexity to the system. Add a shell script to download the data files (similar to the one I pasted above) [0]. Your README file says: {quote} Test Data Files * The system has the ability to check and download required data files from the source, as specified in the properties file. * If you want to manually download files please use the link provided below. {quote} Why are you calling it "test data files"? Also, as a user, I don't fucking care about what ability the system has. I want information on *how to use whatever fucking ability the system has*. You need to be specific in terms of steps the user needs to take. Secondly, "please use the link provided below" is equivalent to telling the user to "go f*** yourself". Just provide an exact command that will enable the user to download those files! Please understand that if someone is not able to use your system due to poor documentation or poor interface, then your system is useless to him/her. As of now, this entire project is useless to anyone except you or me. I urge you to fix this aspect at the earliest. [0] - https://paste.fedoraproject.org/paste/TaYvUCH7vSQRdkYGRkmPnQ > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106233#comment-16106233 ] Ishan Chattopadhyaya commented on SOLR-10317: - And please learn some markdown! Your README file [0] is unreadable due to horizontal scrollers in each section. [0] - https://github.com/viveknarang/lucene-solr/tree/SolrNightlyBenchmarks/dev-tools/SolrNightBenchmarks > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106232#comment-16106232 ] Ishan Chattopadhyaya commented on SOLR-10317: - I saw this in the section for running last n days of commits. {code} * While there is no feature implemented (yet), that supports automatic crunching of performance data for commits for last (n) days, there is still a way to do this. * Prepare and arrange (in order) a list of commit hashes for the (n) days, either one for each day or all of the commits for (n) days. * Now create empty files with name (no extension) EXACTLY equal to commit hashes in the sub-folder /data/commit_queue/ found in the directory as configured in the properties file for benchmarkAppDirectory parameter. The hashes will be picked in the order of their creation (FIFO). * Finally, in the section "Steps to run" follow all the steps but for step 8 above, use parameter --from-queue instead of --latest-commit {code} Just use a shell script like this to do this: {code} for n in {1..30}; do date=`date --date="$n days ago" +%Y-%m-%d`; git log --after="$date 00:00" --before="$date 23:59"|head -1|cut -f 2 -d " "; done {code} bq. Steps to run for N days are added in the README file. Yes, currently, it has to be done manually but I will try to add a new feature so this could be done using parameters very soon. No need to add a parameter. Shell scripting can be your friend. All the steps that you mentioned can be automated using the above script. Kindly update the steps asap. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16106194#comment-16106194 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya] I have fixed the label bug on graphs. Please see [http://212.47.242.214/MergedViewStandalone.html] Regards Vivek > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16104154#comment-16104154 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya] I have added additional capability to archive performance data (--archive) and clear performance data (--clear-data). Please check the latest readme file. Regards Vivek > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16103731#comment-16103731 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya] Steps to run for (n) days are added in the README file. Yes, currently, it has to be done manually but I will try to add a new feature so this could be done using parameters very soon. Menu for standalone and Cloud have been re-arranged as you described above. Now a "More Details" link on each panel on the Unified view pages will take the user to the linked page. The back button is redundant since there is are links in the top menu to browse around as required. If you still prefer it, please let me know and I will add it on the pages. Please check [http://212.47.242.214] I will compile and list the TODO items shortly. Regards Vivek > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16102215#comment-16102215 ] Ishan Chattopadhyaya commented on SOLR-10317: - Thanks Vivek. As per our offline discussion, I can see that you have collapsed the "serial" benchmarks into the "concurrent" benchmarks with a threads=1 line. Looks much less confusing now! Can you please document the steps on how to run the benchmarks for past n days (not past n commits), with just one commit per day? Even if the steps are manual in nature, please document them. Going forward, it would be good to have some parameters to let us do that (perhaps also allow for a time of the day). Also, I suggest that top bar menu be very simple: Standalone Metrics and Cloud Metrics. Both could link to the "merged" view (better call it "unified" view). Each of the graphs can be linked to the individual pages, and each page could have a back button that takes the user back to the "unified" page. Also, can you please list down the major TODO items? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16097386#comment-16097386 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya] Multiple threads have already been implemented. Please see [http://212.47.242.214/IndexingThroughputBenchmarkStandaloneConcurrent.html] I think you got confused between two different tests. One test uses HTTP Solr client while the other with threads uses Concurrent Update client. The ability to locate and download test data files is awesome and now the system will be able to check the data file presence and download the missing file(s) by itself as required. The source URL is configurable through the properties file. Regards Vivek > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16097128#comment-16097128 ] Ishan Chattopadhyaya commented on SOLR-10317: - Also, add something like this [0] to help users download the data files used for benchmarking. [0] - https://paste.fedoraproject.org/paste/TaYvUCH7vSQRdkYGRkmPnQ > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16097127#comment-16097127 ] Ishan Chattopadhyaya commented on SOLR-10317: - For Indexing benchmarks, we need multiple threads. For querying benchmarks, we need multiple threads and the following metrics: Average latency, 90th %ile latency, 99th%ile latency, throughput. Please refer to JMeter for this functionality. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16084044#comment-16084044 ] Vivek Narang commented on SOLR-10317: - Hello [~ichattopadhyaya] Wikipedia data - done. Unified page view - pending. Merging menu items - done. Code formatting - in progress, partially done. Args parameter renaming - done. Log4j - done. Label adjustment - pending. Regards Vivek > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16081756#comment-16081756 ] Ishan Chattopadhyaya commented on SOLR-10317: - [~vivek.nar...@uga.edu], please report on the status of these pending items. https://issues.apache.org/jira/browse/SOLR-10317?focusedCommentId=16066003=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-16066003 I'm mostly concerned about items 1 (using Wikipedia data in your benchmarks) and 4 (having a unified, one page view of your graphs). > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16078068#comment-16078068 ] Vivek Narang commented on SOLR-10317: - Hi, [~michael.sun] Thanks. I will look into it soon. Regards. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16075623#comment-16075623 ] Michael Sun commented on SOLR-10317: [~vivek.nar...@uga.edu] I saw you made some good progress. Not sure what framework you decided to use. But if you wanted to use your own, make sure add CPU util metric in test result. You are welcome to use code in my patch for CPU measurement. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16075617#comment-16075617 ] Michael Sun commented on SOLR-10317: Just uploaded a second cut of Solr benchmark framework I built at work. In addition to object model, this patch has a working JSON facet benchmark using JMeter and measures CPU usage during benchmark. More importantly, the code for JMeter and CPU usage are easily reusable in other benchmarks. As mentioned earlier, the idea of this framework is to avoid building a new framework for all Solr performance work of upcoming years, by carefully designed, extensible object model and reusable components. I am aware of several existing Solr benchmarks, inc. one actively being developed now. Any feedback is appreciated. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16066003#comment-16066003 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya] & [~mkhludnev] Below are the action points (In order of importance), from the Round 1 evaluation meeting. - Switching from test data, currently used, to the use of Wikipedia data for index benchmarking. - Merging menu items (eg. Merging standalone numeric query tests under category Standalone Metrics etc.) - Code formatting (Removing spaces, indentation, missing javadocs etc.) - Adding an option to put all the graps on one page (category wise) - Args parameter renaming. - Use of log4j as applicable. - Adjust labels on graphs/charts I am happy and thankful to have awesome Mentors like you! Regards Vivek Narang > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16057599#comment-16057599 ] Michael Sun commented on SOLR-10317: bq. It changed after identifying and handling resource contention What was the resource contention? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16057279#comment-16057279 ] Vivek Narang commented on SOLR-10317: - Hi [~michael.sun] It changed after identifying and handling resource contention. Sadly for Indexing (using ConcurrentUpdateSolrClient) on the SolrCloud, there are still fluctuations noted (I am guessing because ConcurrentUpdateSolrClient uses HttpSolrClient instead of CloudSolrClient, see: https://github.com/apache/lucene-solr/blob/master/solr/solrj/src/java/org/apache/solr/client/solrj/impl/ConcurrentUpdateSolrClient.java#L105-L107) > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16055839#comment-16055839 ] Michael Sun commented on SOLR-10317: bq.the fluctuations are now been contained [~vivek.nar...@uga.edu] That's cool. What is the change to solve the fluctuations? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16052624#comment-16052624 ] Vivek Narang commented on SOLR-10317: - Hi [~michael.sun]. Like you, I was also deeply concerned about the fluctuations in the resulting metrics and after digging deep into the problem, for last two days, and applying critical fixes, I am happy to inform you that the fluctuations are now been contained. Please access [http://212.47.227.9/prod/]. Some important changes are still remaining though. Regards > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > SOLR-10317.patch, solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16049655#comment-16049655 ] Michael Sun commented on SOLR-10317: Just uploaded the first cut of Solr benchmark I built during my work, as one more option for community for benchmarking. There are a few good benchmarks in the community for different use cases, using different frameworks. The goal of my benchmark, in short, is to design an extensible, standardized benchmark that can be used for a variety of common performance use cases. Nightly performance regression tests are very important. Meanwhile it would be good if we can reuse the same benchmark for capacity planning, scalability study, troubleshooting, etc., which has slightly different requirement to nightly tests. It would be a good saving for everyone in community if he only needs to extend the benchmark, not rebuild one, for his own use cases in near future. In addition, the benchmark includes a variety of instruments to help understand why the performance is, in addition to what the performance is. One obvious reason is that answering why is the primary goal for some use cases, such as troubleshooting, scalability study. Meanwhile it also helps to build 'correct' performance tests. For example, performance bottleneck discovered in tests may not be a code defect but some setup issue. Being able to analyze a bit can make sure the performance tests are testing the right thing. Designing a good benchmark is one of my primary jobs at work. So I will continue to elaborate the framework and add new tests. There are a few good benchmarks for Solr. Also [~vivek.nar...@uga.edu] has done a great job in designing a few new test cases. I can help you in porting or adding new test cases with my framework if you like. The patch mainly includes object model and a sample test to demonstrate object model. More components will follow. It's an option to community of course but I do think community can benefit from this contribution. Any feedback is appreciated. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16045363#comment-16045363 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya] An Update. I have implemented a logic to calculate the Latency (for numeric queries) and corresponding percentiles (75th, 95th, 99th & 99.9th) for Latency. Additionally, I am in the process of adding mechanisms for AND and OR type queries for Numeric fields. These new features are not yet visible on the front end, however, their logic is already added. These new features will appear soon. Regards. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16042085#comment-16042085 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya] An Update. I have added a mechanism with which a set of commits collected in a single day are processed for benchmarks at midnight now. Please view a sample run of this batch mode (that was processed today) here: [http://212.47.227.9/prod/]. The benchmark (bach-mode) cycle is now triggered by Jenkins every midnight [EST]. Regards > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16039986#comment-16039986 ] Michael Sun commented on SOLR-10317: bq. any thoughts, please? Well, this is a good topic. Framework is fun and always a tradeoff. :) On one side, we should not over invest on framework because the tests and numbers produced by tests are things really matter. That's the target we should focus on. On the other hand, if framework is under invested, in long run there is high cost. One way to understand the cost is to look how many Solr performance frameworks there are already, not to mention a few not published (but I personally know). I am one of guys who built my own (going to be open source soon, adding one more into the collection). One of the motivation for me to build one is 'only need to extend framework but not rebuild one for all Solr performance work in near future'. In my option this is important and should be part of the goals of this project. And be part of the discussion about framework choice too (no matter which one is chosen). > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16038504#comment-16038504 ] Ishan Chattopadhyaya commented on SOLR-10317: - bq. The reason behind me extending upon your framework is that it already has many flexible, ready to use resources and that it is written in one language. I am comfortable using one language over two languages together. Thanks for the clarification. Exactly what I was looking for, i.e. *reasons why you chose* the current suite over Shalin's suite. Personally, I don't care whichever suite you are using so long as your stated goals are met, and we achieve parity with whatever prior work exists already, and the overall suite is flexible enough to add more benchmarks later. bq. the closest that I am in making it dynamic and self-dependent is showing relevant commit messages with each metric point http://212.47.227.9/prod/NumericQueryBenchmarkStandalone.html (please hover over any point to see the relevant commit message). Maintaining a separate JSON file containing significant commit and message sounds like a good way forward. Is it too difficult to plot that info on the graphs, like it is done in Shalin's or Mike's graphs? bq. I will try to add a feature through which you would be able to view all the graphs together. Sounds good. Btw, Shalin's suite has got some tests that I didn't see in your suite or proposal. I know that you've plotted some metrics on a per commit basis in a popup window for indexing (memory consumption over the course of indexing), but having an independent graph on GC while indexing and other similar graphs that Shalin has added to that suite would be good. Don't stretch yourself for them right now, but it would be awesome if you can add them at the end of your GSoC project, for the sake of completeness of the suite. [~shalinmangar], [~michael.sun], any thoughts, please? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16038003#comment-16038003 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya] I think there is some confusion and I think I should give some explanation. - I am not building a new framework, but I am extending the benchmarks framework that you created [https://github.com/chatman/solr-upgrade-tests] as mentioned in the proposal. - The reason behind me extending upon your framework is that it already has many flexible, ready to use resources and that it is written in one language. I am comfortable using one language over two languages together. - For the remaining things, I already am in the process of using the required resources from Shalin's work and adding to the framework that you created. - As far as tagging/addition of significant events go while I think that the current logic in Shalin's code base related to listing significant events is hard coded [https://github.com/shalinmangar/solr-perf-tools/blob/master/src/python/bench.py#L32-L87], the closest that I am in making it dynamic and self-dependent is showing relevant commit messages with each metric point [http://212.47.227.9/prod/NumericQueryBenchmarkStandalone.html] (please hover over any point to see the relevant commit message). - I will try to add a feature through which you would be able to view all the graphs together. Please access the latest codebase [https://github.com/viveknarang/lucene-solr/tree/SolrNightlyBenchmarks/dev-tools/SolrNightBenchmarks]. As per the agreement in the proposal, the code for benchmarks suite is under dev-tools framework and in the SolrNightlyBenchmarks branch. Regards > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16036133#comment-16036133 ] Ishan Chattopadhyaya commented on SOLR-10317: - bq. This dataset looks great! Seems like CC0 license, which might be fine. However, before we actually start using it, I'd like to get a clearance from ASF's legal team to make sure we're good. Give me a few days for this. LEGAL-313 > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16036128#comment-16036128 ] Ishan Chattopadhyaya commented on SOLR-10317: - I'm very confused. Are you planning to use Shalin's benchmarking utility and extend upon it? If no, why not? Please understand that Shalin or I are NOT asking you to use one platform/utility or the other, but simply asking you for a clarification on the motivations behind *why* you are building a new platform you are using instead of something that already exists. So far, I just see no clear answer, just beating around the bush. In terms of the differences, two good things I like about the reports that Shalin's suite generates (https://issues.apache.org/jira/secure/attachment/12843260/indexing.html) are: # all graphs can be viewed at once # supports tagging/addition of significant events If you choose not to use Shalin's suite, can you please address these two issues in the new suite that you are building? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16036110#comment-16036110 ] Vivek Narang commented on SOLR-10317: - Hi [~ichattopadhyaya] I have restored Jenkins on the new server please access [http://212.47.227.9/prod]. At this moment you may not see any metric data but soon the numbers should become visible. Regards. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16033983#comment-16033983 ] Ishan Chattopadhyaya commented on SOLR-10317: - bq. Hello Ishan Chattopadhyaya, I have added a mechanism with which the failed tests sessions are identified and handled. The running processes, if any, (solr/zookeeper) from the last failed session are located and destroyed, any files/folders created during the failed session are located and removed and Metric data files are only updated with new statistics when the test session is completely successful. This mechanism will make this suite more self-sufficient. Sounds good. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16033978#comment-16033978 ] Ishan Chattopadhyaya commented on SOLR-10317: - bq. I was, however, trying to understand the IndexThreads class that you have written. Can you please explain its working when you have some free time? I've used IndexThreads in another project, and here's the reference: https://github.com/lintool/IR-Reproducibility/blob/master/systems/lucene/ingester/src/main/java/luceneingester/IndexThreads.java I remember getting that from Lucene's benchmark module, but I cannot find it now. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16033906#comment-16033906 ] Vivek Narang commented on SOLR-10317: - Hello [~ichattopadhyaya], I have added a mechanism with which the failed tests sessions are identified and handled. The running processes, if any, (solr/zookeeper) from the last failed session are located and destroyed, any files/folders created during the failed session are located and removed and Metric data files are only updated with new statistics when the test session is completely successful. This mechanism will make this suite more self-sufficient. Please access the latest code at [https://github.com/viveknarang/lucene-solr/tree/SolrNightlyBenchmarks/dev-tools/SolrNightBenchmarks]. Regards. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16033224#comment-16033224 ] Vivek Narang commented on SOLR-10317: - Hi [~shalinmangar] Thanks for your suggestions above. Your code for indexing is excellent and I will be using that in the project. I was, however, trying to understand the IndexThreads class that you have written. Can you please explain its working when you have some free time? Thanks in advance. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16029990#comment-16029990 ] Vivek Narang commented on SOLR-10317: - Hello, I have been down with illness for last three days, I will resume my activity shortly. Regards > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16027329#comment-16027329 ] Ishan Chattopadhyaya commented on SOLR-10317: - bq. In fact it's one of the reasons that I suggested to start a scoping doc to start conversation early on. There exists one: https://issues.apache.org/jira/secure/attachment/12861725/Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16027030#comment-16027030 ] Michael Sun commented on SOLR-10317: bq. motivation behind creating yet another benchmarking utility That's a good question. In fact it's one of the reasons that I suggested to start a scoping doc to start conversation early on. (https://issues.apache.org/jira/browse/SOLR-10317?focusedCommentId=16011107=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-16011107) Here are my two cents for a few areas that can be improved in addition to increasing test coverage. [~vivek.nar...@uga.edu] can articulate. 1. Currently benchmark tells us how Solr perform but it can also help to tell why Solr perform in this way. A good example of effort in this direction is the telemetry (https://esrally.readthedocs.io/en/latest/telemetry.html) in rally framework. 2. Provide baseline data for capacity planning. For capacity planning, it requires some data such as CPU, disk etc. for specific workloads and benchmark can provide that. 3. Extensibility: a benchmark can be easily extended to include new components. For example, JMeter can be a good load generator for scalability study for Solr cluster with hundreds of nodes and it should be easy to extend current test case to use JMeter to replace existing load generator. This may require an object model at different abstraction level compared to existing benchmarks. 4. Support more Solr setup and data type. For example, wiki data is a good but tweets data can be better in studying Solr performance for real time analytics use cases. 5. Last but not least, as any engineering tool, I was hoping the benchmark suite can standardize Solr performance effort, promote code reuse and facilitate collaboration. This requires good understanding for all use cases and careful design. Of course, this doesn't need to be all done for GoC project. Not to scare [~vivek.nar...@uga.edu] :) Overall, this project is a good initiative and a good venue to continue this discussion. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16026567#comment-16026567 ] Ishan Chattopadhyaya commented on SOLR-10317: - Can you please explain the test methodology for measuring QPS? What is the field type (Trie or Point / Int or Double or Long or Float)? Firing the same query again and again is useless. Is that what you're trying? Also, what is the latency? bq. Why is the motivation behind creating yet another benchmarking utility? Can you please answer this ^ ? bq. I think the central question is why you chose one over the other? Can you please answer this ^ ? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16026549#comment-16026549 ] Vivek Narang commented on SOLR-10317: - Hello [~ichattopadhyaya], After getting inspired by source code in SolrMeter I have come up with the logic to do a kind of a stress test where I am estimating the QPS for a set of numeric queries (as an example for now.). Please access [http://212.47.227.9/dev/NumericQueryBenchmarkCloud.html]. I am observing a strange co-incidence - The QPS measured for a query looking for a specific number is lowest and the QPS measured for a query looking for all those numbers greater than a number. Is there an explanation for this? QPS(Field:Number) < QPS (Field:[Number1 TO Number2]) < QPS(Field:[* TO Number]) < QPS(Field:[Number TO *]) This has been observed over a set of commits and not one commit. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16023721#comment-16023721 ] Jan Høydahl commented on SOLR-10317: We could also draw inspiration or code from https://github.com/elastic/rally which is really easy to use and quite extensible. No idea how much effort a solr-rally port would take or if it would be legally viable? :-) The tool supports some nice data sets ootb, see https://esrally.readthedocs.io/en/latest/race.html. A side effect would thus be that we could run same benchmarks to compare against ES as well :-) > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16023607#comment-16023607 ] Shalin Shekhar Mangar commented on SOLR-10317: -- bq. I would love to know your thoughts as to how we can come up with a better and much more comprehensive suite, based on the prior work done The current solr-perf-tools has support for: # Indexing JSON on a single node solr instance with schemaless configs # Indexing wiki-1kb docs on single node solr instance with fixed schema # Indexing wiki-4kb docs on single node solr instance with fixed schema # Indexing wiki-1kb docs on 2 shard, 1 replica solr cloud See the report attached at SOLR-9863. I like wiki data because the Lucene benchmarks also use it (but not exactly the same data) which gives us a sense how much overhead Solr has over Lucene. I also spent more time on non-cloud single node benchmarks because those are easier to reason about and debug. Troubleshooting cloud performance problems is much more difficult without establishing a baseline using consistent single node benchmarks. We can go two ways from here: # Cleanup/refactor the code to make the tool easier to extend and add benchmarks e.g. instead of writing python code for testing variants of a test, perhaps a test description written in json or a DSL could be executed # Forget about the cleaning the code and just add more benchmarks both indexing and query and not worry about code duplication There are arguments for both e.g. #1 above will encourage more people to contribute to benchmarks but #2 above will make your progress faster. As for the benchmarks themselves, we already have basic indexing benchmarks there so we need to get started with query benchmarks. There are just so many possibilities here but we can start with uncached query performance first. For this you need to extract terms out of your data set, classify them according to frequency and test all combinations on a solr instance with query/filter cache disabled but ensure that we graph them separately. Reuse the indexes built by the indexing test. As an example, see the BooleanQuery section at https://home.apache.org/~mikemccand/lucenebench/ and the extracted terms data at https://github.com/mikemccand/luceneutil/blob/master/tasks/wikimedium.10M.tasks. Then repeat this with using both {{q}} and {{fq}} params. Then with dismax query parser and so on. Then iterate again with caches enabled this time. Then repeat with re-indexing data during the query tests (both cached and uncached cases). Use your imagination. Focus on correctness and repeatability. > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail:
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16023442#comment-16023442 ] Ishan Chattopadhyaya commented on SOLR-10317: - bq. Hi Shalin Shekhar Mangar Thank you for sharing the link to your prior work. I briefly went through your work before, using the link to your work that was mentioned in the description section above. I think the purpose of this initiative is to have a comprehensive benchmarking suite based on the prior work done in this area and addition of many other features/tests that helps the Solr community. I would love to know your thoughts as to how we can come up with a better and much more comprehensive suite, based on the prior work done. Regards. There were some projects that I listed in the description, including Shalin's and yours. I think the central question is why you chose one over the other? > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
[jira] [Commented] (SOLR-10317) Solr Nightly Benchmarks
[ https://issues.apache.org/jira/browse/SOLR-10317?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16023438#comment-16023438 ] Ishan Chattopadhyaya commented on SOLR-10317: - bq. Hi, I am trying to find a way to measure the QPS for query benchmarking. The way I understand it is that it is a way to measure the querying capacity for a Solr node or a cluster. Can anyone recommend me the correct mechanism of estimating QPS? Thanks for the help. Please look at SolrMeter. https://github.com/tflobbe/solrmeter > Solr Nightly Benchmarks > --- > > Key: SOLR-10317 > URL: https://issues.apache.org/jira/browse/SOLR-10317 > Project: Solr > Issue Type: Task >Reporter: Ishan Chattopadhyaya > Labels: gsoc2017, mentor > Attachments: changes-lucene-20160907.json, > changes-solr-20160907.json, managed-schema, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks.docx, > Narang-Vivek-SOLR-10317-Solr-Nightly-Benchmarks-FINAL-PROPOSAL.pdf, > solrconfig.xml > > > Solr needs nightly benchmarks reporting. Similar Lucene benchmarks can be > found here, https://home.apache.org/~mikemccand/lucenebench/. > Preferably, we need: > # A suite of benchmarks that build Solr from a commit point, start Solr > nodes, both in SolrCloud and standalone mode, and record timing information > of various operations like indexing, querying, faceting, grouping, > replication etc. > # It should be possible to run them either as an independent suite or as a > Jenkins job, and we should be able to report timings as graphs (Jenkins has > some charting plugins). > # The code should eventually be integrated in the Solr codebase, so that it > never goes out of date. > There is some prior work / discussion: > # https://github.com/shalinmangar/solr-perf-tools (Shalin) > # https://github.com/chatman/solr-upgrade-tests/blob/master/BENCHMARKS.md > (Ishan/Vivek) > # SOLR-2646 & SOLR-9863 (Mark Miller) > # https://home.apache.org/~mikemccand/lucenebench/ (Mike McCandless) > # https://github.com/lucidworks/solr-scale-tk (Tim Potter) > There is support for building, starting, indexing/querying and stopping Solr > in some of these frameworks above. However, the benchmarks run are very > limited. Any of these can be a starting point, or a new framework can as well > be used. The motivation is to be able to cover every functionality of Solr > with a corresponding benchmark that is run every night. > Proposing this as a GSoC 2017 project. I'm willing to mentor, and I'm sure > [~shalinmangar] and [~markrmil...@gmail.com] would help here. -- This message was sent by Atlassian JIRA (v6.3.15#6346) - To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org