[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=297222&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-297222 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 19/Aug/19 15:44 Start Date: 19/Aug/19 15:44 Worklog Time Spent: 10m Work Description: asfgit commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 297222) Time Spent: 18h 40m (was: 18.5h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 18h 40m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian Jira (v8.3.2#803003)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=296165&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-296165 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Aug/19 08:56 Start Date: 16/Aug/19 08:56 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-521938303 @michaelandrepearce squashed commits and pushed. BTW, I added some code to avoid file leak in race conditions in PageSubscriptionImpl::internalGetNext() : ``` serverMessage = ((PageReader) cache).getMessage(retPos, false, true); PageCache previousPageCache = pageReaders.putIfAbsent(retPos.getPageNr(), (PageReader) cache); if (previousPageCache != null && previousPageCache != cache) { // Maybe other cursor iterators have added page reader, we have to close this one to avoid file leak cache.close(); } ``` This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 296165) Time Spent: 18.5h (was: 18h 20m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 18.5h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=296163&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-296163 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Aug/19 08:54 Start Date: 16/Aug/19 08:54 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-521938303 @michaelandrepearce squashed commits and pushed. BTW, I added some code to avoid file leak in race conditions in PageSubscriptionImpl::internalGetNext() : ` serverMessage = ((PageReader) cache).getMessage(retPos, false, true); PageCache previousPageCache = pageReaders.putIfAbsent(retPos.getPageNr(), (PageReader) cache); if (previousPageCache != null && previousPageCache != cache) { // Maybe other cursor iterators have added page reader, we have to close this one to avoid file leak cache.close(); }` This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 296163) Time Spent: 18h 10m (was: 18h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 18h 10m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=296164&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-296164 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Aug/19 08:54 Start Date: 16/Aug/19 08:54 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-521938303 @michaelandrepearce squashed commits and pushed. BTW, I added some code to avoid file leak in race conditions in PageSubscriptionImpl::internalGetNext() : `serverMessage = ((PageReader) cache).getMessage(retPos, false, true); PageCache previousPageCache = pageReaders.putIfAbsent(retPos.getPageNr(), (PageReader) cache); if (previousPageCache != null && previousPageCache != cache) { // Maybe other cursor iterators have added page reader, we have to close this one to avoid file leak cache.close(); }` This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 296164) Time Spent: 18h 20m (was: 18h 10m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 18h 20m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=296161&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-296161 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Aug/19 08:53 Start Date: 16/Aug/19 08:53 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-521938303 @michaelandrepearce squashed commits and pushed. BTW, I added some code to avoid file leak in race conditions in PageSubscriptionImpl::internalGetNext() : ` serverMessage = ((PageReader) cache).getMessage(retPos, false, true); PageCache previousPageCache = pageReaders.putIfAbsent(retPos.getPageNr(), (PageReader) cache); if (previousPageCache != null && previousPageCache != cache) { // Maybe other cursor iterators have added page reader, we have to close this one to avoid file leak cache.close(); }` This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 296161) Time Spent: 17h 50m (was: 17h 40m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 17h 50m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=296162&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-296162 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Aug/19 08:53 Start Date: 16/Aug/19 08:53 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-521938303 @michaelandrepearce squashed commits and pushed. BTW, I added some code to avoid file leak in race conditions in PageSubscriptionImpl::internalGetNext() : ` serverMessage = ((PageReader) cache).getMessage(retPos, false, true); PageCache previousPageCache = pageReaders.putIfAbsent(retPos.getPageNr(), (PageReader) cache); if (previousPageCache != null && previousPageCache != cache) { // Maybe other cursor iterators have added page reader, we have to close this one to avoid file leak cache.close(); } ` This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 296162) Time Spent: 18h (was: 17h 50m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 18h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=296160&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-296160 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Aug/19 08:52 Start Date: 16/Aug/19 08:52 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-521938303 @michaelandrepearce squashed commits and pushed. BTW, I added some code to avoid file leak in race conditions in PageSubscriptionImpl::internalGetNext() : ` serverMessage = ((PageReader) cache).getMessage(retPos, false, true); PageCache previousPageCache = pageReaders.putIfAbsent(retPos.getPageNr(), (PageReader) cache); if (previousPageCache != null && previousPageCache != cache) { // Maybe other cursor iterators have added page reader, we have to close this one to avoid file leak cache.close(); }` This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 296160) Time Spent: 17h 40m (was: 17.5h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 17h 40m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=296134&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-296134 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Aug/19 08:20 Start Date: 16/Aug/19 08:20 Worklog Time Spent: 10m Work Description: michaelandrepearce commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-521928878 @wy96f can you squash commits pls This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 296134) Time Spent: 17.5h (was: 17h 20m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 17.5h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=296129&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-296129 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Aug/19 07:52 Start Date: 16/Aug/19 07:52 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-521921421 pushed commit to close page reader when cursor iterator is closed This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 296129) Time Spent: 17h 20m (was: 17h 10m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 17h 20m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=296074&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-296074 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Aug/19 04:56 Start Date: 16/Aug/19 04:56 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-521883667 @michaelandrepearce Can you wait a moment? When the browse cursor iterators(like the one created by QueueControl::browse, browse only consumer, etc) are closed, the page opened by them is not closed. I would push another commit to fix this. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 296074) Time Spent: 17h 10m (was: 17h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 17h 10m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=296071&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-296071 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Aug/19 04:48 Start Date: 16/Aug/19 04:48 Worklog Time Spent: 10m Work Description: michaelandrepearce commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-521882489 @wy96f can you squash commits so we can merge? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 296071) Time Spent: 17h (was: 16h 50m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 17h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=295238&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-295238 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Aug/19 06:30 Start Date: 15/Aug/19 06:30 Worklog Time Spent: 10m Work Description: michaelandrepearce commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-521531183 @franz1981 @wy96f @clebertsuconic as no further comments/feedback will merge this tomorrow. Last call :) This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 295238) Time Spent: 16h 50m (was: 16h 40m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 16h 50m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=290612&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-290612 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 07/Aug/19 17:22 Start Date: 07/Aug/19 17:22 Worklog Time Spent: 10m Work Description: michaelandrepearce commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-519191794 @franz1981 @clebertsuconic @wy96f is this ready to merge? Would be good to get this body of work into the next release cutting this week This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 290612) Time Spent: 16.5h (was: 16h 20m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 16.5h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=290614&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-290614 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 07/Aug/19 17:22 Start Date: 07/Aug/19 17:22 Worklog Time Spent: 10m Work Description: michaelandrepearce commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-519192081 All tests ive run so far dont have any regressions This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 290614) Time Spent: 16h 40m (was: 16.5h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 16h 40m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=289506&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-289506 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 06/Aug/19 07:59 Start Date: 06/Aug/19 07:59 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518556136 @franz1981 > And while catching up, the producers are not seriously decreasing their rate? This PR has helped in your tests? This PR helped to improve consumer throughput while entering into page mode with lots of subscribers. I didn't see the producers rate seriously decreased, although producers slowed down bcs consumers used extra io. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 289506) Time Spent: 16h 20m (was: 16h 10m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 16h 20m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288873&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288873 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 11:31 Start Date: 05/Aug/19 11:31 Worklog Time Spent: 10m Work Description: franz1981 commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518197382 @wy96f > it's possible the pages are constantly dropping in an out while consumers are cathing up, resulting in performance drop Yep, but that would affect both consumer(s) and producers, not just producers: anyway I'm struggling to reproduce it reliably, so probably will dig into it much further before checking with this PR :+1: > In our test consumers were falling behind producers all the time. I saw ~80GB page files left after test lasting 1 hour. And while catching up, the producers are not seriously decreasing their rate? This PR has helped in your tests? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 288873) Time Spent: 16h 10m (was: 16h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 16h 10m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288790&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288790 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 08:32 Start Date: 05/Aug/19 08:32 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518139732 @franz1981 In our test consumers were falling behind producers all the time. I saw ~80GB page files left after test lasting 1 hour. I'm not sure whether it's related to this pr. Considering 10GB page data produced, it's possible the pages are constantly dropping in an out while consumers are cathing up, resulting in performance drop. Could you try this pr to check it? :) This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 288790) Time Spent: 16h (was: 15h 50m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 16h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288756&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288756 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 08:00 Start Date: 05/Aug/19 08:00 Worklog Time Spent: 10m Work Description: franz1981 commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518129069 > Be great to have a recreator. Will work on provide one soon > Also if in master already its unrelated to the work here (eg this work causes no regression in perf). I would like to avoid conflating issues That's a good point: although I've hoped that would be kinda related to the work done here, but I see that is a different scenario...will move the discuttion into the dev list, thanks :+1: This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 288756) Time Spent: 15h 50m (was: 15h 40m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 15h 50m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288753&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288753 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 07:53 Start Date: 05/Aug/19 07:53 Worklog Time Spent: 10m Work Description: michaelandrepearce commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518126703 are you sure its not underlying disk/infra issue, im very surprised by scenario you explain. Be great to have a recreator. Also if in master already its unrelated to the work here (eg this work causes no regression in perf). I would like to avoid conflating issues This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 288753) Time Spent: 15h 40m (was: 15.5h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 15h 40m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288751&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288751 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 07:52 Start Date: 05/Aug/19 07:52 Worklog Time Spent: 10m Work Description: michaelandrepearce commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518126703 are you sure its not underlying disk/infra issue, im very surprised by scenario you explain. Be great to have a recreator. Also if in master already its unrelated to the work here. I would like to avoid conflating issues This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 288751) Time Spent: 15.5h (was: 15h 20m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 15.5h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288745&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288745 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 07:44 Start Date: 05/Aug/19 07:44 Worklog Time Spent: 10m Work Description: franz1981 commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518124124 @michaelandrepearce I'm using the latest master: my concern is that during `PagingStoreImpl::page` we perform both `storageManagerLock.readLock` and `PagingStoreImpl.lock.writeLock::lock`...so maybe there are consuming operations that prevent these locks to be released This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 288745) Time Spent: 15h 10m (was: 15h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 15h 10m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288746&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288746 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 07:44 Start Date: 05/Aug/19 07:44 Worklog Time Spent: 10m Work Description: franz1981 commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518124124 @michaelandrepearce I'm using the latest master: my concern is that during `PagingStoreImpl::page` we perform both `storageManagerLock.readLock::lock` and `PagingStoreImpl.lock.writeLock::lock`...so maybe there are consuming operations that prevent these locks to be released This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 288746) Time Spent: 15h 20m (was: 15h 10m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 15h 20m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288724&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288724 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 07:16 Start Date: 05/Aug/19 07:16 Worklog Time Spent: 10m Work Description: michaelandrepearce commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518116101 @franz1981 what version was that in. I know alot of locking changes to improve that had been made already to alleviate that. Do you get that still in latest? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 288724) Time Spent: 15h (was: 14h 50m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 15h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288714&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288714 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 07:10 Start Date: 05/Aug/19 07:10 Worklog Time Spent: 10m Work Description: franz1981 commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518114338 @michaelandrepearce > which test? Just a specific scenario I've been fallen into recently, to summarize it: - a topic with many publishers sending durable messages - many durable subscribers - 1 of the subscribers disconnect for some time - publishers start paging - after 10 GB of paged data, the subscriber come back online again - while it catches up, the other publishers/subscribers rate are seriously affected (approaching to 1/2 msg/sec) - when it catches up, things come back normal for everyone This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 288714) Time Spent: 14h 50m (was: 14h 40m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 14h 50m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288712&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288712 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 07:04 Start Date: 05/Aug/19 07:04 Worklog Time Spent: 10m Work Description: michaelandrepearce commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518112452 @franz1981 which test? I know when we page producers slow down this is because of the extra io as goes to paging not memory, which is side effect of paging, and ideally why its good to avoid. Just want to make sure its not that This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 288712) Time Spent: 14h 40m (was: 14.5h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 14h 40m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288711&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288711 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 07:03 Start Date: 05/Aug/19 07:03 Worklog Time Spent: 10m Work Description: michaelandrepearce commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518112452 @franz1981 which test? I know when we page full stop producers slow down this is because of the extra io as goes to paging not memory. Just want to make sure its not that This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 288711) Time Spent: 14.5h (was: 14h 20m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 14.5h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288698&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288698 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 06:57 Start Date: 05/Aug/19 06:57 Worklog Time Spent: 10m Work Description: franz1981 commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518110518 @wy96f A quick question: in your tests have you ever experienced a certain degree of unfairness when consumers are catching up with publishers while consuming a big backlog? I have some test with durable subscribers that shows publishers getting considerably slowed down while consumer(s) are catching up... This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 288698) Time Spent: 14h 20m (was: 14h 10m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 14h 20m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288660&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288660 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 02:44 Start Date: 05/Aug/19 02:44 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518064779 I've added a new commit to reduce file open/close. The log before this commit showed for each queue depage was followed by two deliver after a few seconds, something like this: `2019-07-31 11:24:12,915 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub86 is 4470916 with maxSize = 5242880. Depaged 2000 messages, pendingDelivery=0, intermediateMessageReferences= 2000, queueDelivering=984 2019-07-31 11:24:15,583 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub86, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@2c21ad1e doing deliver. messageReferences=0 2019-07-31 11:24:22,596 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub86, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@2c21ad1e doing deliver. messageReferences=2 2019-07-31 11:24:22,620 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub86 is 4490706 with maxSize = 5242880. Depaged 2000 messages, pendingDelivery=0, intermediateMessageReferences= 2000, queueDelivering=990 2019-07-31 11:24:25,620 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub86, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@2c21ad1e doing deliver. messageReferences=0 2019-07-31 11:24:33,095 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub86, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@2c21ad1e doing deliver. messageReferences=2` `2019-07-31 11:20:34,082 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub49 is 4473860 with maxSize = 5242880. Depaged 2000 messages, pendingDelivery=0, intermediateMessageReferences= 2000, queueDelivering=982 2019-07-31 11:20:34,161 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub49, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@41a2d788 doing deliver. messageReferences=0 2019-07-31 11:20:44,535 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub49, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@41a2d788 doing deliver. messageReferences=2 2019-07-31 11:20:44,802 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub49 is 4473010 with maxSize = 5242880. Depaged 2000 messages, pendingDelivery=0, intermediateMessageReferences= 2000, queueDelivering=971 2019-07-31 11:20:44,999 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub49, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@41a2d788 doing deliver. messageReferences=0 2019-07-31 11:20:55,230 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub49, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@41a2d788 doing deliver. messageReferences=2` It is likely paged message is cleared when delivering a few seconds after depage causing opening/closing file to read message. After this commit, depage is immediately followed by deliver. The log was like this: `2019-08-02 16:35:11,317 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub72 is 2227083 with maxSize = 5242880. Depaged 509 messages, pendingDelivery=491, intermediateMessageReferences= 509, queueDelivering=387 2019-08-02 16:35:11,317 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub72, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=e0d3e34f-b4df-11e9-956a-fa163e6757d7], temp=false]@36c11804 doing deliver. messageReferences=491 2019-08-02 16:35:16,467 DEBUG [org.apache.a
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288659&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288659 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 02:38 Start Date: 05/Aug/19 02:38 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518064779 I've added a new commit to reduce file open/close. The log before this commit showed for each queue depage was followed by two deliver after a few seconds, something like this: `2019-07-31 11:24:12,915 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub86 is 4470916 with maxSize = 5242880. Depaged 2000 messages, pendingDelivery=0, intermediateMessageReferences= 2000, queueDelivering=984 2019-07-31 11:24:15,583 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub86, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@2c21ad1e doing deliver. messageReferences=0 2019-07-31 11:24:22,596 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub86, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@2c21ad1e doing deliver. messageReferences=2 2019-07-31 11:24:22,620 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub86 is 4490706 with maxSize = 5242880. Depaged 2000 messages, pendingDelivery=0, intermediateMessageReferences= 2000, queueDelivering=990 2019-07-31 11:24:25,620 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub86, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@2c21ad1e doing deliver. messageReferences=0 2019-07-31 11:24:33,095 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub86, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@2c21ad1e doing deliver. messageReferences=2` `2019-07-31 11:20:34,082 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub49 is 4473860 with maxSize = 5242880. Depaged 2000 messages, pendingDelivery=0, intermediateMessageReferences= 2000, queueDelivering=982 2019-07-31 11:20:34,161 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub49, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@41a2d788 doing deliver. messageReferences=0 2019-07-31 11:20:44,535 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub49, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@41a2d788 doing deliver. messageReferences=2 2019-07-31 11:20:44,802 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub49 is 4473010 with maxSize = 5242880. Depaged 2000 messages, pendingDelivery=0, intermediateMessageReferences= 2000, queueDelivering=971 2019-07-31 11:20:44,999 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub49, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@41a2d788 doing deliver. messageReferences=0 2019-07-31 11:20:55,230 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub49, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@41a2d788 doing deliver. messageReferences=2` It is likely paged message is cleared when delivering a few seconds after depage causing opening/closing file to read message. After this commit, depage is immediately followed by deliver. The log was like this: `2019-08-02 16:35:11,317 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub72 is 2227083 with maxSize = 5242880. Depaged 509 messages, pendingDelivery=491, intermediateMessageReferences= 509, queueDelivering=387 2019-08-02 16:35:11,317 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub72, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=e0d3e34f-b4df-11e9-956a-fa163e6757d7], temp=false]@36c11804 doing deliver. messageReferences=491 2019-08-02 16:35:16,467 DEBUG [org.apache.a
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=288658&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-288658 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 05/Aug/19 02:36 Start Date: 05/Aug/19 02:36 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-518064779 I've added a new commit to reduce file open/close. The log before this commit showed for each queue depage was followed by two deliver after a few seconds, something like this: `2019-07-31 11:24:12,915 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub86 is 4470916 with maxSize = 5242880. Depaged 2000 messages, pendingDelivery=0, intermediateMessageReferences= 2000, queueDelivering=984 2019-07-31 11:24:15,583 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub86, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@2c21ad1e doing deliver. messageReferences=0 2019-07-31 11:24:22,596 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub86, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@2c21ad1e doing deliver. messageReferences=2 2019-07-31 11:24:22,620 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub86 is 4490706 with maxSize = 5242880. Depaged 2000 messages, pendingDelivery=0, intermediateMessageReferences= 2000, queueDelivering=990 2019-07-31 11:24:25,620 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub86, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@2c21ad1e doing deliver. messageReferences=0 2019-07-31 11:24:33,095 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub86, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@2c21ad1e doing deliver. messageReferences=2` `2019-07-31 11:20:34,082 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub49 is 4473860 with maxSize = 5242880. Depaged 2000 messages, pendingDelivery=0, intermediateMessageReferences= 2000, queueDelivering=982 2019-07-31 11:20:34,161 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub49, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@41a2d788 doing deliver. messageReferences=0 2019-07-31 11:20:44,535 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub49, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@41a2d788 doing deliver. messageReferences=2 2019-07-31 11:20:44,802 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub49 is 4473010 with maxSize = 5242880. Depaged 2000 messages, pendingDelivery=0, intermediateMessageReferences= 2000, queueDelivering=971 2019-07-31 11:20:44,999 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub49, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@41a2d788 doing deliver. messageReferences=0 2019-07-31 11:20:55,230 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub49, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=5d09e339-b32d-11e9-b1ac-fa163e6757d7], temp=false]@41a2d788 doing deliver. messageReferences=2` It is likely paged message is cleared when delivering a few seconds after depage causing opening/closing file to read message. After this commit, depage is immediately followed by deliver. The log was like this: `2019-08-02 16:35:11,317 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Memory Size after depage on queue=sub72 is 2227083 with maxSize = 5242880. Depaged 509 messages, pendingDelivery=491, intermediateMessageReferences= 509, queueDelivering=387 2019-08-02 16:35:11,317 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] QueueImpl[name=sub72, postOffice=PostOfficeImpl [server=ActiveMQServerImpl::serverUUID=e0d3e34f-b4df-11e9-956a-fa163e6757d7], temp=false]@36c11804 doing deliver. messageReferences=491 2019-08-02 16:35:16,467 DEBUG [org.apache.activemq.artemis.core.server.impl.QueueImpl] Queue Me
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=285672&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-285672 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 31/Jul/19 11:19 Start Date: 31/Jul/19 11:19 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-516807021 Pushed to change PageReader per queue. The perf result is below: 1. Running in 51MB size page and 1 page cache in the case of 100 multicast queues. 19000 msg/s received and 9500 msg/s sent vs. 16000 msg/s received and 11000 msg/s sent with previous PageReader cache. consumer tps: ![per_queue_multi_consumer](https://user-images.githubusercontent.com/7719761/62190465-fdc38080-b3a3-11e9-8c25-078fa75e54d2.png) producer tps: ![per_queue_multi_producer](https://user-images.githubusercontent.com/7719761/62190509-159b0480-b3a4-11e9-9cee-5abc42dacb06.png) 2. Running in 5MB size page and 100 page cache in the case of 100 multicast queues. 16000msg/s received and 8200 msg/s sent vs. 15000 msg/s received and 8500 msg/s sent with previous PageReader cache. consumer tps: ![per_queue_multi_consumer_small](https://user-images.githubusercontent.com/7719761/62206438-67ec1d80-b3c4-11e9-842e-463e2a325187.png) producer tps: ![per_queue_multi_producer_small](https://user-images.githubusercontent.com/7719761/62206486-8225fb80-b3c4-11e9-93f4-a8e5e97dbba8.png) 3. Running in 51MB size page and 1 page cache in the case of 1 queue. 16000 msg/s received and 29000 msg/s sent vs. 13500 msg/s received and 3 msg/s sent with previous PageReader cache. consumer tps: ![per_queue_single_consumer](https://user-images.githubusercontent.com/7719761/62190825-dc16c900-b3a4-11e9-94ee-1946b851adea.png) producer tps: ![per_queue_single_producer](https://user-images.githubusercontent.com/7719761/62190864-f0f35c80-b3a4-11e9-932c-f706d569d56e.png) Generally speaking, the consumer tps increase a little and producer tps reduce less, with total tps increasing by about 1000. It's to be noted that there was frequent file opening/closing in second test scenario(5MB size page and 100 page cache). I added some logging and saw two subsequent deliver was scheduled a few seconds after depage. In the case that message of PagedReference was cleared, calling getPriority/checkExpired in deliver() would trigger reading message. Due to the small page file, reading messages were more likely to hit page before the current cursor. Given PageReader before current cursor was closed, this would cause constantly page file open/close(log showed approximately 6000 times per second in average). In first test scenario(51MB size page and 1 page cache), approximately 700 times per second in average of file open/closing occurred. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 285672) Time Spent: 13h 40m (was: 13.5h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 13h 40m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=283969&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-283969 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 29/Jul/19 01:15 Start Date: 29/Jul/19 01:15 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-515812679 I've submitted a new pr #2769 to fix it. After reviewed and merged, I would rebase and push the new code with PageReader per queue. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 283969) Time Spent: 13.5h (was: 13h 20m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 13.5h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=282715&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-282715 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 25/Jul/19 15:44 Start Date: 25/Jul/19 15:44 Worklog Time Spent: 10m Work Description: clebertsuconic commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-515097267 If it's an easy fix, you should fix it first on a separate PR. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 282715) Time Spent: 13h 20m (was: 13h 10m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 13h 20m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=282440&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-282440 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 25/Jul/19 05:51 Start Date: 25/Jul/19 05:51 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-514906870 I would use separate pr and as we haven't released yet. Even use that original jira to fix #2740 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 282440) Time Spent: 13h 10m (was: 13h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 13h 10m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=282405&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-282405 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 25/Jul/19 02:57 Start Date: 25/Jul/19 02:57 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-514875375 > I will need some time to review this last week. > > Can you rebase again after I merged your change? I'm working on PageReader per queue these days. Unluckily some problem was encountered. After investigation, I found it was related to #2740 which would sync files while closing. Given we put PageReader in PageSubscription meaning 100 times of file open/close for each page in our case(100 queues), although page file is only opened for reading, the sync operation does cause great performance loss - ~5000 msg/s received and 13000 msg/s sent now compared with 16000 msg/s received and 11000 msg/s sent of pr. Should I file a new jira and pr to fix the problem introuduced by #2740 or just fix it in this one? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 282405) Time Spent: 13h (was: 12h 50m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 13h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=282396&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-282396 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 25/Jul/19 02:29 Start Date: 25/Jul/19 02:29 Worklog Time Spent: 10m Work Description: clebertsuconic commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-514870176 I will need some time to review this last week. Can you rebase again after I merged your change? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 282396) Time Spent: 12h 50m (was: 12h 40m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 12h 50m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=280302&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-280302 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 22/Jul/19 09:57 Start Date: 22/Jul/19 09:57 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-513725688 Rebased and fixed to add deps and configuration This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 280302) Time Spent: 12h 40m (was: 12.5h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 12h 40m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=279042&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-279042 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 18/Jul/19 14:37 Start Date: 18/Jul/19 14:37 Worklog Time Spent: 10m Work Description: clebertsuconic commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r304953204 ## File path: artemis-server/pom.xml ## @@ -89,6 +89,10 @@ ${project.version} test + Review comment: You have to add: ``xml mvn:org.jctools/jctools-core/${jctools.version} ``` as a dependency on artemis-core. see my other comment This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 279042) Time Spent: 12h 20m (was: 12h 10m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 12h 20m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=279043&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-279043 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 18/Jul/19 14:37 Start Date: 18/Jul/19 14:37 Worklog Time Spent: 10m Work Description: clebertsuconic commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r304953204 ## File path: artemis-server/pom.xml ## @@ -89,6 +89,10 @@ ${project.version} test + Review comment: You have to add: ```xml mvn:org.jctools/jctools-core/${jctools.version} ``` as a dependency on artemis-core. see my other comment This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 279043) Time Spent: 12.5h (was: 12h 20m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 12.5h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=279040&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-279040 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 18/Jul/19 14:36 Start Date: 18/Jul/19 14:36 Worklog Time Spent: 10m Work Description: clebertsuconic commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-512843045 Can you rebase this and fix the merging issue? I have merged your other PR and there's a clash now. In the process, please add the new component into features as a dependency. on features.xml ```xml artemis-common scr mvn:org.apache.activemq/artemis-features/${pom.version}/cfg mvn:org.apache.activemq/artemis-features/${pom.version}/xml/artemis mvn:org.apache.geronimo.specs/geronimo-jms_2.0_spec/${geronimo.jms.2.spec.version} mvn:org.apache.commons/commons-configuration2/${commons.config.version} mvn:org.apache.commons/commons-text/1.6 mvn:org.apache.commons/commons-lang3/${commons.lang.version} mvn:org.jctools/jctools-core/${jctools.version} mvn:org.apache.activemq/activemq-artemis-native/${activemq-artemis-native-version} mvn:org.apache.activemq/artemis-server-osgi/${pom.version} ``` This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 279040) Time Spent: 12h 10m (was: 12h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 12h 10m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277928&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277928 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 17/Jul/19 02:06 Start Date: 17/Jul/19 02:06 Worklog Time Spent: 10m Work Description: clebertsuconic commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r304193664 ## File path: artemis-server/pom.xml ## @@ -89,6 +89,10 @@ ${project.version} test + Review comment: I had to add this on features on my branch to be able to run tests. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277928) Time Spent: 12h (was: 11h 50m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 12h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277916&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277916 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 17/Jul/19 01:30 Start Date: 17/Jul/19 01:30 Worklog Time Spent: 10m Work Description: wy96f commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r304187715 ## File path: artemis-server/src/main/java/org/apache/activemq/artemis/core/server/impl/QueueImpl.java ## @@ -324,13 +325,7 @@ public String debug() { out.println("consumer: " + holder.consumer.debug()); } - for (MessageReference reference : intermediateMessageReferences) { - out.print("Intermediate reference:" + reference); - } - - if (intermediateMessageReferences.isEmpty()) { - out.println("No intermediate references"); - } + out.println("Intermediate reference size is " + intermediateMessageReferences.size()); Review comment: MpscUnboundedArrayQueue might perform well in journal mode where there are many producers and one consumer. The main reason I used MpscUnboundedArrayQueue is I need to calculate size of messages both in intermediateMessageReferences and messageReferences, i.e.`intermediateMessageReferences.size() + messageReferences.size() < MAX_DEPAGE_NUM` in QueueImpl::needsDepage. However time complexity of size() for ConcurrentLinkedQueue is O(n), so I replaced it with MpscUnboundedArrayQueue(I did't find other queues fit for our scenario). I didn't think about issues like kafraf, debug by introducing this. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277916) Time Spent: 11h 50m (was: 11h 40m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 11h 50m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277338&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277338 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 08:57 Start Date: 16/Jul/19 08:57 Worklog Time Spent: 10m Work Description: franz1981 commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303792800 ## File path: artemis-server/src/main/java/org/apache/activemq/artemis/core/server/impl/QueueImpl.java ## @@ -324,13 +325,7 @@ public String debug() { out.println("consumer: " + holder.consumer.debug()); } - for (MessageReference reference : intermediateMessageReferences) { - out.print("Intermediate reference:" + reference); - } - - if (intermediateMessageReferences.isEmpty()) { - out.println("No intermediate references"); - } + out.println("Intermediate reference size is " + intermediateMessageReferences.size()); Review comment: Not so easy I think...for bounded versions of qs yes (already supported in the current/maybe next release), but for unbounded is not that easy... We are sure we really use iterator? or is iterator + clear at the end? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277338) Time Spent: 11h 40m (was: 11.5h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 11h 40m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277337&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277337 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 08:56 Start Date: 16/Jul/19 08:56 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303794549 ## File path: artemis-server/src/main/java/org/apache/activemq/artemis/core/server/impl/QueueImpl.java ## @@ -324,13 +325,7 @@ public String debug() { out.println("consumer: " + holder.consumer.debug()); } - for (MessageReference reference : intermediateMessageReferences) { - out.print("Intermediate reference:" + reference); - } - - if (intermediateMessageReferences.isEmpty()) { - out.println("No intermediate references"); - } + out.println("Intermediate reference size is " + intermediateMessageReferences.size()); Review comment: So debug like this only make it into code if it was needed at some point to debug an operational issue, as such i would suggest it was needed. My question is how much perf improvement is actually gained from this one particular change. I would expect the real perf was in the change in paging logic not the change of queue, which to me anyhow seems a little unrelated to improving performance for lots of subscribers (the jira) because its actually anyhow on the ingress to the core queue, not the consumption. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277337) Time Spent: 11.5h (was: 11h 20m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 11.5h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277336&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277336 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 08:55 Start Date: 16/Jul/19 08:55 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303794549 ## File path: artemis-server/src/main/java/org/apache/activemq/artemis/core/server/impl/QueueImpl.java ## @@ -324,13 +325,7 @@ public String debug() { out.println("consumer: " + holder.consumer.debug()); } - for (MessageReference reference : intermediateMessageReferences) { - out.print("Intermediate reference:" + reference); - } - - if (intermediateMessageReferences.isEmpty()) { - out.println("No intermediate references"); - } + out.println("Intermediate reference size is " + intermediateMessageReferences.size()); Review comment: So debug like this only make it into code if it was needed at some point to debug an operational issue, as such i would suggest it was needed. My question is how much perf improvement is actually gained from this one particular change. I would expect the real perf was in the change in paging logic not the change of queue, which to me anyhow seems a little unrelated to improving performance for lots of subscribers because its actually anyhow on the ingress to the core queue, not the consumption. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277336) Time Spent: 11h 20m (was: 11h 10m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 11h 20m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277333&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277333 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 08:51 Start Date: 16/Jul/19 08:51 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303794549 ## File path: artemis-server/src/main/java/org/apache/activemq/artemis/core/server/impl/QueueImpl.java ## @@ -324,13 +325,7 @@ public String debug() { out.println("consumer: " + holder.consumer.debug()); } - for (MessageReference reference : intermediateMessageReferences) { - out.print("Intermediate reference:" + reference); - } - - if (intermediateMessageReferences.isEmpty()) { - out.println("No intermediate references"); - } + out.println("Intermediate reference size is " + intermediateMessageReferences.size()); Review comment: So debug like this only make it into code if it was needed at some point to debug an operational issue, as such i would suggest it was needed. My question is how much perf improvement is actually gained from this one particular change. I would expect the real perf was in the change in paging logic not the change of queue. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277333) Time Spent: 11h 10m (was: 11h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 11h 10m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277330&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277330 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 08:47 Start Date: 16/Jul/19 08:47 Worklog Time Spent: 10m Work Description: franz1981 commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303792800 ## File path: artemis-server/src/main/java/org/apache/activemq/artemis/core/server/impl/QueueImpl.java ## @@ -324,13 +325,7 @@ public String debug() { out.println("consumer: " + holder.consumer.debug()); } - for (MessageReference reference : intermediateMessageReferences) { - out.print("Intermediate reference:" + reference); - } - - if (intermediateMessageReferences.isEmpty()) { - out.println("No intermediate references"); - } + out.println("Intermediate reference size is " + intermediateMessageReferences.size()); Review comment: Not so easy I think...for bounded versions of qs yet (already supported in the current/maybe next release), but for unbounded is not that easy... We are sure we really use iterator? or is iterator + clear at the end? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277330) Time Spent: 11h (was: 10h 50m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 11h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277327&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277327 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 08:40 Start Date: 16/Jul/19 08:40 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303789825 ## File path: artemis-server/pom.xml ## @@ -89,6 +89,10 @@ ${project.version} test + Review comment: Need to ensure this is updated in the OSGI / Karaf bundles config to ensure its bundled and loaded correctly, else those running in Karaf / OSGI they will get issue. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277327) Time Spent: 10h 50m (was: 10h 40m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 10h 50m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277326&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277326 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 08:39 Start Date: 16/Jul/19 08:39 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303789264 ## File path: artemis-server/src/main/java/org/apache/activemq/artemis/core/server/impl/QueueImpl.java ## @@ -324,13 +325,7 @@ public String debug() { out.println("consumer: " + holder.consumer.debug()); } - for (MessageReference reference : intermediateMessageReferences) { - out.print("Intermediate reference:" + reference); - } - - if (intermediateMessageReferences.isEmpty()) { - out.println("No intermediate references"); - } + out.println("Intermediate reference size is " + intermediateMessageReferences.size()); Review comment: @franz1981 is this relatively trivial? Could always copy the class in temporarily and add the feature in artemis copy as an interim. (also would avoid a karaf issue with extra deps) This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277326) Time Spent: 10h 40m (was: 10.5h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 10h 40m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277324&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277324 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 08:35 Start Date: 16/Jul/19 08:35 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303787724 ## File path: artemis-server/src/main/java/org/apache/activemq/artemis/core/server/impl/QueueImpl.java ## @@ -324,13 +325,7 @@ public String debug() { out.println("consumer: " + holder.consumer.debug()); } - for (MessageReference reference : intermediateMessageReferences) { - out.print("Intermediate reference:" + reference); - } - - if (intermediateMessageReferences.isEmpty()) { - out.println("No intermediate references"); - } + out.println("Intermediate reference size is " + intermediateMessageReferences.size()); Review comment: Do we need MpscUnboundedArrayQueue then in this PR to make it work? Im conscious of removing useful debugging features. Perf shouldnt be at the cost of feature / operational need. Can always be re-added back once it supports? In a later PR perf improvement. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277324) Time Spent: 10.5h (was: 10h 20m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 10.5h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277323&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277323 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 08:34 Start Date: 16/Jul/19 08:34 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-511720732 @wy96f that makes it alot easier to understand the graphs :) thanks. +1 on perf improvements This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277323) Time Spent: 10h 20m (was: 10h 10m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 10h 20m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277319&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277319 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 08:31 Start Date: 16/Jul/19 08:31 Worklog Time Spent: 10m Work Description: franz1981 commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303785677 ## File path: tests/smoke-tests/src/test/java/org/apache/activemq/artemis/tests/smoke/jmx/JmxConnectionTest.java ## @@ -27,8 +27,8 @@ import java.rmi.server.RemoteRef; import io.netty.util.internal.PlatformDependent; -import io.netty.util.internal.shaded.org.jctools.util.UnsafeAccess; import org.apache.activemq.artemis.tests.smoke.common.SmokeTestBase; +import org.jctools.util.UnsafeAccess; Review comment: We are working on a VarHandle version and we already allows different AtomicXXX versions to be used (with a small perf drop): probably java 12 is fine too, but I haven't tested it personally... The problem will be when some Unsafe specific methods will be scattered around VarHandle/Other classes and will be officially deleted from Unsafe... This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277319) Time Spent: 9h 50m (was: 9h 40m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 9h 50m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277321&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277321 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 08:32 Start Date: 16/Jul/19 08:32 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-511720535 @wy96f i was thinking simply at just a broker level would be enough for me, just as a simple enable/kill switch (toggle) for the feature. If you did that you could even make it default on, just leave the option for people to disable should they need. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277321) Time Spent: 10h 10m (was: 10h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 10h 10m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277320&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277320 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 08:32 Start Date: 16/Jul/19 08:32 Worklog Time Spent: 10m Work Description: franz1981 commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303786131 ## File path: artemis-server/src/main/java/org/apache/activemq/artemis/core/server/impl/QueueImpl.java ## @@ -324,13 +325,7 @@ public String debug() { out.println("consumer: " + holder.consumer.debug()); } - for (MessageReference reference : intermediateMessageReferences) { - out.print("Intermediate reference:" + reference); - } - - if (intermediateMessageReferences.isEmpty()) { - out.println("No intermediate references"); - } + out.println("Intermediate reference size is " + intermediateMessageReferences.size()); Review comment: I will take a note to implement one on JCTools :+1: This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277320) Time Spent: 10h (was: 9h 50m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 10h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277318&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277318 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 08:28 Start Date: 16/Jul/19 08:28 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303784485 ## File path: tests/smoke-tests/src/test/java/org/apache/activemq/artemis/tests/smoke/jmx/JmxConnectionTest.java ## @@ -27,8 +27,8 @@ import java.rmi.server.RemoteRef; import io.netty.util.internal.PlatformDependent; -import io.netty.util.internal.shaded.org.jctools.util.UnsafeAccess; import org.apache.activemq.artemis.tests.smoke.common.SmokeTestBase; +import org.jctools.util.UnsafeAccess; Review comment: you say until 11...does that mean after its knackered? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277318) Time Spent: 9h 40m (was: 9.5h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 9h 40m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277313&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277313 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 08:25 Start Date: 16/Jul/19 08:25 Worklog Time Spent: 10m Work Description: wy96f commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303783168 ## File path: artemis-server/src/main/java/org/apache/activemq/artemis/core/server/impl/QueueImpl.java ## @@ -324,13 +325,7 @@ public String debug() { out.println("consumer: " + holder.consumer.debug()); } - for (MessageReference reference : intermediateMessageReferences) { - out.print("Intermediate reference:" + reference); - } - - if (intermediateMessageReferences.isEmpty()) { - out.println("No intermediate references"); - } + out.println("Intermediate reference size is " + intermediateMessageReferences.size()); Review comment: MpscUnboundedArrayQueue doesn't support iterator, so I had to remove it. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277313) Time Spent: 9.5h (was: 9h 20m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 9.5h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277305&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277305 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 16/Jul/19 07:58 Start Date: 16/Jul/19 07:58 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-511398767 @michaelandrepearce I deployed the test without MAX_DEPAGE_NUM limit during depage. The performance result as follows: 1. Running in 51MB size page and 1 page cache in the case of 100 multicast queues. The consumer tps dropped to ~11000 compared to ~16000 in pr. consumer tps for pr: ![ppos_multi_consumer](https://user-images.githubusercontent.com/7719761/61275838-5c540080-a7e1-11e9-961c-b405bd481f87.png) producer tps for pr: ![ppos_multi_producer](https://user-images.githubusercontent.com/7719761/61275853-6544d200-a7e1-11e9-8f4d-88456ed7ebb8.png) consumer tps for no MAX_DEPAGE_NUM limit: ![ppos_multi_consumer_nolimit](https://user-images.githubusercontent.com/7719761/61275891-7857a200-a7e1-11e9-969a-5692d6f2f02a.png) producer tps for no MAX_DEPAGE_NUM limit: ![ppos_multi_producer_nolimit](https://user-images.githubusercontent.com/7719761/61275911-84436400-a7e1-11e9-8051-dd429ae7f6d8.png) 2. Running in 5MB size page and 100 page cache in the case of 100 multicast queues. The consumer tps dropped to ~1 compared to ~15000 in pr. consumer tps for pr: ![ppos_multi_consumer_small_page](https://user-images.githubusercontent.com/7719761/61276001-c076c480-a7e1-11e9-8ddf-dbb05515999c.png) producer tps for pr: ![ppos_multi_producer_small_page](https://user-images.githubusercontent.com/7719761/61276024-cbc9f000-a7e1-11e9-8acc-fec8c1a76206.png) consumer tps for no MAX_DEPAGE_NUM limit: ![ppos_multi_consumer_small_page_nolimit](https://user-images.githubusercontent.com/7719761/61276040-d97f7580-a7e1-11e9-8f8c-cf7ab646d969.png) producer tps for no MAX_DEPAGE_NUM limit: ![ppos_multi_producer_small_page_nolimit](https://user-images.githubusercontent.com/7719761/61276064-e1d7b080-a7e1-11e9-80e0-91885b9642d4.png) 3. Running in 51MB size page and 1 page cache in the case of 1 queue. Performance results similar. consumer tps for pr: ![ppos_single_consumer](https://user-images.githubusercontent.com/7719761/61276355-7215f580-a7e2-11e9-8d4f-eb95bb4362f9.png) producer tps for pr: ![ppos_single_producer](https://user-images.githubusercontent.com/7719761/61276376-7d692100-a7e2-11e9-9f23-3fc140625abd.png) consumer tps for no MAX_DEPAGE_NUM limit: ![ppos_single_consumer_nolimit](https://user-images.githubusercontent.com/7719761/61276386-83f79880-a7e2-11e9-8242-964cf1753bb8.png) producer tps for no MAX_DEPAGE_NUM limit: ![ppos_single_producer_nolimit](https://user-images.githubusercontent.com/7719761/61276400-8bb73d00-a7e2-11e9-8862-5856128f2c21.png) Seems like paged message removed and re reading file resulted in the performance drop particularly in the case of heavy usage of memory(test 1,2). Even if there was enough free memory and messages were not gced(test 3), the performance didn't make a big difference - instead memory footprint might be increased. Considering we've totally relied on the kernel page cache, it seems not to need spare messages in memory. I personally assume this is an import feature. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277305) Time Spent: 9h 20m (was: 9h 10m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 9h 20m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 q
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277034&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277034 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 22:40 Start Date: 15/Jul/19 22:40 Worklog Time Spent: 10m Work Description: clebertsuconic commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-511596184 I just came back from vacations.. I will need another day to look into this. Please wait for me, let me review it.. don't merge it yet :) This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277034) Time Spent: 9h 10m (was: 9h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 9h 10m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277032&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277032 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 22:35 Start Date: 15/Jul/19 22:35 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-511594857 re test 3, it might be the graphs size, but on squinting it does seems a significant perf differece, from what i can make 15k to 30k, which one is this PR and which one is master? And same with latency, seems the bottom is 18 where top is 40, that also significant change. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277032) Time Spent: 9h (was: 8h 50m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 9h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277031&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277031 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 22:34 Start Date: 15/Jul/19 22:34 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-511594857 re test 3, it might be the graphs size, but on squiting it does seems a significant perf differece, from what i can make 15k to 30k, which one is this PR and which one is master? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277031) Time Spent: 8h 50m (was: 8h 40m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 8h 50m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277030&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277030 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 22:31 Start Date: 15/Jul/19 22:31 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-511594112 Forgot to say, great stuff as always!! Personally prefer this impl to the other. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277030) Time Spent: 8h 40m (was: 8.5h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 8h 40m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277027&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277027 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 22:28 Start Date: 15/Jul/19 22:28 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303662881 ## File path: artemis-server/src/main/java/org/apache/activemq/artemis/core/paging/cursor/impl/PageReader.java ## @@ -0,0 +1,88 @@ +/** + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.activemq.artemis.core.paging.cursor.impl; + +import org.apache.activemq.artemis.core.paging.PagedMessage; +import org.apache.activemq.artemis.core.paging.cursor.PageCache; +import org.apache.activemq.artemis.core.paging.cursor.PagePosition; +import org.apache.activemq.artemis.core.paging.impl.Page; +import org.jboss.logging.Logger; + +public class PageReader implements PageCache { + private static final Logger logger = Logger.getLogger(PageReader.class); + + private Page page; + + public PageReader(Page page) { + this.page = page; + } + + @Override + public long getPageId() { + return page.getPageId(); + } + + @Override + public int getNumberOfMessages() { + return page.getNumberOfMessages(); + } + + @Override + public void setMessages(PagedMessage[] messages) { Review comment: ditto on comment below for this one too This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277027) Time Spent: 8h 10m (was: 8h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 8h 10m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277028&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277028 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 22:28 Start Date: 15/Jul/19 22:28 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303663059 ## File path: artemis-server/src/main/java/org/apache/activemq/artemis/core/paging/cursor/impl/PageReader.java ## @@ -0,0 +1,88 @@ +/** + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.activemq.artemis.core.paging.cursor.impl; + +import org.apache.activemq.artemis.core.paging.PagedMessage; +import org.apache.activemq.artemis.core.paging.cursor.PageCache; +import org.apache.activemq.artemis.core.paging.cursor.PagePosition; +import org.apache.activemq.artemis.core.paging.impl.Page; +import org.jboss.logging.Logger; + +public class PageReader implements PageCache { + private static final Logger logger = Logger.getLogger(PageReader.class); + + private Page page; + + public PageReader(Page page) { + this.page = page; + } + + @Override + public long getPageId() { + return page.getPageId(); + } + + @Override + public int getNumberOfMessages() { + return page.getNumberOfMessages(); + } + + @Override + public void setMessages(PagedMessage[] messages) { + } + + @Override + public PagedMessage[] getMessages() { + throw new UnsupportedOperationException("get messages should not be called for index page cache"); + } + + @Override + public boolean isLive() { + return false; + } + + @Override + public synchronized PagedMessage getMessage(PagePosition pagePosition) { + if (pagePosition.getMessageNr() >= getNumberOfMessages()) { + return null; + } + try { + PagedMessage msg; + if (pagePosition.getFileOffset() != -1) { +msg = page.readMessage(pagePosition.getFileOffset(), pagePosition.getMessageNr(), pagePosition.getMessageNr(), pagePosition); + } else { +if (logger.isTraceEnabled()) { + logger.trace("get message from pos " + pagePosition, new Exception("trace get message")); +} +msg = page.readMessage(0, 0, pagePosition.getMessageNr(), pagePosition); + } + return msg; + } catch (Exception e) { + throw new RuntimeException(e.getMessage(), e); + } + } + + @Override + public synchronized void close() { + try { + page.close(false); + } catch (Exception e) { + logger.warn("Closing page " + page.getPageId() + " occurs exception:", e); Review comment: should be a parametised error with an error code. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277028) Time Spent: 8h 20m (was: 8h 10m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 8h 20m > Remaining Estimate: 0h > > We noticed that there was a s
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277029&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277029 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 22:28 Start Date: 15/Jul/19 22:28 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303663059 ## File path: artemis-server/src/main/java/org/apache/activemq/artemis/core/paging/cursor/impl/PageReader.java ## @@ -0,0 +1,88 @@ +/** + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.activemq.artemis.core.paging.cursor.impl; + +import org.apache.activemq.artemis.core.paging.PagedMessage; +import org.apache.activemq.artemis.core.paging.cursor.PageCache; +import org.apache.activemq.artemis.core.paging.cursor.PagePosition; +import org.apache.activemq.artemis.core.paging.impl.Page; +import org.jboss.logging.Logger; + +public class PageReader implements PageCache { + private static final Logger logger = Logger.getLogger(PageReader.class); + + private Page page; + + public PageReader(Page page) { + this.page = page; + } + + @Override + public long getPageId() { + return page.getPageId(); + } + + @Override + public int getNumberOfMessages() { + return page.getNumberOfMessages(); + } + + @Override + public void setMessages(PagedMessage[] messages) { + } + + @Override + public PagedMessage[] getMessages() { + throw new UnsupportedOperationException("get messages should not be called for index page cache"); + } + + @Override + public boolean isLive() { + return false; + } + + @Override + public synchronized PagedMessage getMessage(PagePosition pagePosition) { + if (pagePosition.getMessageNr() >= getNumberOfMessages()) { + return null; + } + try { + PagedMessage msg; + if (pagePosition.getFileOffset() != -1) { +msg = page.readMessage(pagePosition.getFileOffset(), pagePosition.getMessageNr(), pagePosition.getMessageNr(), pagePosition); + } else { +if (logger.isTraceEnabled()) { + logger.trace("get message from pos " + pagePosition, new Exception("trace get message")); +} +msg = page.readMessage(0, 0, pagePosition.getMessageNr(), pagePosition); + } + return msg; + } catch (Exception e) { + throw new RuntimeException(e.getMessage(), e); + } + } + + @Override + public synchronized void close() { + try { + page.close(false); + } catch (Exception e) { + logger.warn("Closing page " + page.getPageId() + " occurs exception:", e); Review comment: should be a parametised log error with an error code. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277029) Time Spent: 8.5h (was: 8h 20m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 8.5h > Remaining Estimate: 0h > > We noticed that there was a s
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277026&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277026 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 22:27 Start Date: 15/Jul/19 22:27 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303662809 ## File path: artemis-server/src/main/java/org/apache/activemq/artemis/core/paging/cursor/impl/PageReader.java ## @@ -0,0 +1,88 @@ +/** + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.activemq.artemis.core.paging.cursor.impl; + +import org.apache.activemq.artemis.core.paging.PagedMessage; +import org.apache.activemq.artemis.core.paging.cursor.PageCache; +import org.apache.activemq.artemis.core.paging.cursor.PagePosition; +import org.apache.activemq.artemis.core.paging.impl.Page; +import org.jboss.logging.Logger; + +public class PageReader implements PageCache { + private static final Logger logger = Logger.getLogger(PageReader.class); + + private Page page; + + public PageReader(Page page) { + this.page = page; + } + + @Override + public long getPageId() { + return page.getPageId(); + } + + @Override + public int getNumberOfMessages() { + return page.getNumberOfMessages(); + } + + @Override + public void setMessages(PagedMessage[] messages) { + } + + @Override + public PagedMessage[] getMessages() { Review comment: would this not be dangerous? either we have code relying on this, and need to support it, or we should remove it from the interface/api so people dont code against it. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277026) Time Spent: 8h (was: 7h 50m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 8h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277025&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277025 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 22:23 Start Date: 15/Jul/19 22:23 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303660459 ## File path: tests/smoke-tests/src/test/java/org/apache/activemq/artemis/tests/smoke/jmx/JmxConnectionTest.java ## @@ -27,8 +27,8 @@ import java.rmi.server.RemoteRef; import io.netty.util.internal.PlatformDependent; -import io.netty.util.internal.shaded.org.jctools.util.UnsafeAccess; import org.apache.activemq.artemis.tests.smoke.common.SmokeTestBase; +import org.jctools.util.UnsafeAccess; Review comment: @franz1981 what occurs post java 8 here? Same with any internal Unsafe access in the MPSC Queue used , just worried we tune for java 8 where unsafe is king, and then find on java 11 perf is lost, and we rely on a new lib that isnt needed.. It might be all good, i just want to play dumb and make sure, and i know jctools is a bit of your area of expertise, and will know the answer outright, and just tell me its no worries. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277025) Time Spent: 7h 50m (was: 7h 40m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 7h 50m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277023&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277023 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 22:19 Start Date: 15/Jul/19 22:19 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303660459 ## File path: tests/smoke-tests/src/test/java/org/apache/activemq/artemis/tests/smoke/jmx/JmxConnectionTest.java ## @@ -27,8 +27,8 @@ import java.rmi.server.RemoteRef; import io.netty.util.internal.PlatformDependent; -import io.netty.util.internal.shaded.org.jctools.util.UnsafeAccess; import org.apache.activemq.artemis.tests.smoke.common.SmokeTestBase; +import org.jctools.util.UnsafeAccess; Review comment: @franz1981 what occurs post java 8 here? Same with any internal Unsafe access in the MPSC Queue used , just worried we tune for java 8 where unsafe is king, and then find on java 11 perf is lost, and we rely on a new lib that isnt needed.. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277023) Time Spent: 7h 40m (was: 7.5h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 7h 40m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277022&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277022 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 22:18 Start Date: 15/Jul/19 22:18 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#discussion_r303660459 ## File path: tests/smoke-tests/src/test/java/org/apache/activemq/artemis/tests/smoke/jmx/JmxConnectionTest.java ## @@ -27,8 +27,8 @@ import java.rmi.server.RemoteRef; import io.netty.util.internal.PlatformDependent; -import io.netty.util.internal.shaded.org.jctools.util.UnsafeAccess; import org.apache.activemq.artemis.tests.smoke.common.SmokeTestBase; +import org.jctools.util.UnsafeAccess; Review comment: @franz1981 what occurs post java 8 here? Same with any internal Unsafe access in the MPSC Queue used This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277022) Time Spent: 7.5h (was: 7h 20m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 7.5h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=277020&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-277020 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 22:11 Start Date: 15/Jul/19 22:11 Worklog Time Spent: 10m Work Description: michaelpearce-gain commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-511587691 @wy96f any chance for higher resolution images? (and whilst at it english titles to help understand the graphs). Re impl detail, As i mentioned in the email lists, i think what ever change is made, as this is soo core to foundation of the broker, it will be a big and risky change perf wise, it is key (a must for me) this is something we let people toggle on, and we leave the default behaviour as was so user dont get suprises and also de-risks, at least for a few releases, this way if there is a perf issue or usecase not covered, there wont be any impact to other users. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 277020) Time Spent: 7h 20m (was: 7h 10m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 7h 20m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=276712&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-276712 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 13:26 Start Date: 15/Jul/19 13:26 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-511398767 @michaelandrepearce I deployed the test without MAX_DEPAGE_NUM limit during depage. The performance result as follows: 1. Running in 51MB size page and 1 page cache in the case of 100 multicast queues. The consumer tps dropped to ~11000 compared to ~16000 in pr. ![image](https://user-images.githubusercontent.com/7719761/61217429-45120600-a742-11e9-8f52-835b10cbabaf.png) ![image](https://user-images.githubusercontent.com/7719761/61217451-4f340480-a742-11e9-8d5a-690b69312d6f.png) 2. Running in 5MB size page and 100 page cache in the case of 100 multicast queues. The consumer tps dropped to ~1 compared to ~15000 in pr. ![image](https://user-images.githubusercontent.com/7719761/61217513-6e329680-a742-11e9-991e-b90a83f2001e.png) ![image](https://user-images.githubusercontent.com/7719761/61217537-74c10e00-a742-11e9-84d1-23c136eaa59a.png) 3. Running in 51MB size page and 1 page cache in the case of 1 queue. Performance results similar. ![image](https://user-images.githubusercontent.com/7719761/61217894-4abc1b80-a743-11e9-88ad-c9fa9a1f2a61.png) ![image](https://user-images.githubusercontent.com/7719761/61217904-514a9300-a743-11e9-9334-47725ba33773.png) Seems like paged message removed and re reading file resulted in the performance drop particularly in the case of heavy usage of memory(test 1,2). Even if there was enough free memory and messages were not gced(test 3), the performance didn't make a big difference - instead memory footprint might be increased. Considering we've totally relied on the kernel page cache, it seems not to need spare messages in memory. I personally assume this is an import feature. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 276712) Time Spent: 7h 10m (was: 7h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 7h 10m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=276706&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-276706 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 13:22 Start Date: 15/Jul/19 13:22 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2730: ARTEMIS-2399 Fix performance degradation when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2730#issuecomment-511387411 closing this as we use another approach. see #2750 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 276706) Time Spent: 7h (was: 6h 50m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 7h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=276704&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-276704 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 13:20 Start Date: 15/Jul/19 13:20 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750#issuecomment-511398767 @michaelandrepearce I deployed the test without MAX_DEPAGE_NUM limit during depage. The performance result as follows: 1. Running in 51MB size page and 1 page cache in the case of 100 multicast queues. The consumer tps dropped to ~11000 compared to ~16000 in pr. ![image](https://user-images.githubusercontent.com/7719761/61217429-45120600-a742-11e9-8f52-835b10cbabaf.png) ![image](https://user-images.githubusercontent.com/7719761/61217451-4f340480-a742-11e9-8d5a-690b69312d6f.png) 2. Running in 5MB size page and 100 page cache in the case of 100 multicast queues. The consumer tps dropped to ~1 compared to ~15000 in pr. ![image](https://user-images.githubusercontent.com/7719761/61217513-6e329680-a742-11e9-991e-b90a83f2001e.png) ![image](https://user-images.githubusercontent.com/7719761/61217537-74c10e00-a742-11e9-84d1-23c136eaa59a.png) 3. Running in 51MB size page and 1 page cache in the case of 1 queue. Performance results similar. ![image](https://user-images.githubusercontent.com/7719761/61217894-4abc1b80-a743-11e9-88ad-c9fa9a1f2a61.png) ![image](https://user-images.githubusercontent.com/7719761/61217904-514a9300-a743-11e9-9334-47725ba33773.png) Seems like paged message removed and re reading file resulted in the performance drop particularly in the case of heavy usage of memory(test #1,#2). Even if there was enough free memory and messages were not gced(test #3), the performance didn't make a big difference - instead memory footprint might be increased. Considering we've totally relied on the kernel page cache, it seems not to need spare messages in memory. I personally assume this is an import feature. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 276704) Time Spent: 6h 50m (was: 6h 40m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 6h 50m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=276679&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-276679 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 12:46 Start Date: 15/Jul/19 12:46 Worklog Time Spent: 10m Work Description: wy96f commented on pull request #2730: ARTEMIS-2399 Fix performance degradation when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2730 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 276679) Time Spent: 6h 40m (was: 6.5h) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 6h 40m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=276678&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-276678 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 12:46 Start Date: 15/Jul/19 12:46 Worklog Time Spent: 10m Work Description: wy96f commented on issue #2730: ARTEMIS-2399 Fix performance degradation when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2730#issuecomment-511387411 #2750 This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 276678) Time Spent: 6.5h (was: 6h 20m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 6.5h > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)
[jira] [Work logged] (ARTEMIS-2399) Improve performance when there are a lot of subscribers
[ https://issues.apache.org/jira/browse/ARTEMIS-2399?focusedWorklogId=276676&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-276676 ] ASF GitHub Bot logged work on ARTEMIS-2399: --- Author: ASF GitHub Bot Created on: 15/Jul/19 12:45 Start Date: 15/Jul/19 12:45 Worklog Time Spent: 10m Work Description: wy96f commented on pull request #2750: ARTEMIS-2399 Improve performance when there are a lot of subscribers URL: https://github.com/apache/activemq-artemis/pull/2750 Based on discussion from [http://activemq.2283324.n4.nabble.com/Improve-paging-performance-when-there-are-lots-of-subscribers-td4751333.html](url) I decided to take the approach suggested by franz which is different than the one before. Now file offset is put in the PagePostion for reading directly from file and the extra index cache layer is removed. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org Issue Time Tracking --- Worklog Id: (was: 276676) Time Spent: 6h 20m (was: 6h 10m) > Improve performance when there are a lot of subscribers > --- > > Key: ARTEMIS-2399 > URL: https://issues.apache.org/jira/browse/ARTEMIS-2399 > Project: ActiveMQ Artemis > Issue Type: Improvement > Components: Broker >Affects Versions: 2.9.0 > Environment: broker 2.9.0 > cpu: 4 cores, memory: 8G, disk: ssd 500G > broker.xml: > 60 > > 51Mb > 50Mb > 1 > PAGE > > -1 >Reporter: yangwei >Priority: Major > Time Spent: 6h 20m > Remaining Estimate: 0h > > We noticed that there was a significant drop in performance when entering > page mode in the case of multiple subscribers. > We created a topic and 100 queues bound to it. We ran our _GrinderRunner > test_ in our inner test infra cluster with 500 threads producing message and > 560 threads, each one picked a random queue to subscribe. The test showed > performance is bad: 13000 msg/s sent and 5000 msg/s received. -- This message was sent by Atlassian JIRA (v7.6.14#76016)