[jira] [Updated] (HDFS-12482) Provide a configuration to adjust the weight of EC recovery tasks to adjust the speed of recovery
[ https://issues.apache.org/jira/browse/HDFS-12482?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lei (Eddy) Xu updated HDFS-12482: - Resolution: Fixed Fix Version/s: 3.0.0 Status: Resolved (was: Patch Available) Thanks for the reviews [~xiaochen]. Committed to trunk and 3.0.0. > Provide a configuration to adjust the weight of EC recovery tasks to adjust > the speed of recovery > - > > Key: HDFS-12482 > URL: https://issues.apache.org/jira/browse/HDFS-12482 > Project: Hadoop HDFS > Issue Type: Improvement > Components: erasure-coding >Affects Versions: 3.0.0-alpha4 >Reporter: Lei (Eddy) Xu >Assignee: Lei (Eddy) Xu >Priority: Minor > Labels: hdfs-ec-3.0-nice-to-have > Fix For: 3.0.0 > > Attachments: HDFS-12482.00.patch, HDFS-12482.01.patch, > HDFS-12482.02.patch, HDFS-12482.03.patch, HDFS-12482.04.patch, > HDFS-12482.05.patch > > > The relative speed of EC recovery comparing to 3x replica recovery is a > function of (EC codec, number of sources, NIC speed, and CPU speed, and etc). > Currently the EC recovery has a fixed {{xmitsInProgress}} of {{max(# of > sources, # of targets)}} comparing to {{1}} for 3x replica recovery, and NN > uses {{xmitsInProgress}} to decide how much recovery tasks to schedule to the > DataNode this we can add a coefficient for user to tune the weight of EC > recovery tasks. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-12482) Provide a configuration to adjust the weight of EC recovery tasks to adjust the speed of recovery
[ https://issues.apache.org/jira/browse/HDFS-12482?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lei (Eddy) Xu updated HDFS-12482: - Attachment: HDFS-12482.05.patch Update to fix checkstyle warnings. > Provide a configuration to adjust the weight of EC recovery tasks to adjust > the speed of recovery > - > > Key: HDFS-12482 > URL: https://issues.apache.org/jira/browse/HDFS-12482 > Project: Hadoop HDFS > Issue Type: Improvement > Components: erasure-coding >Affects Versions: 3.0.0-alpha4 >Reporter: Lei (Eddy) Xu >Assignee: Lei (Eddy) Xu >Priority: Minor > Labels: hdfs-ec-3.0-nice-to-have > Attachments: HDFS-12482.00.patch, HDFS-12482.01.patch, > HDFS-12482.02.patch, HDFS-12482.03.patch, HDFS-12482.04.patch, > HDFS-12482.05.patch > > > The relative speed of EC recovery comparing to 3x replica recovery is a > function of (EC codec, number of sources, NIC speed, and CPU speed, and etc). > Currently the EC recovery has a fixed {{xmitsInProgress}} of {{max(# of > sources, # of targets)}} comparing to {{1}} for 3x replica recovery, and NN > uses {{xmitsInProgress}} to decide how much recovery tasks to schedule to the > DataNode this we can add a coefficient for user to tune the weight of EC > recovery tasks. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-12482) Provide a configuration to adjust the weight of EC recovery tasks to adjust the speed of recovery
[ https://issues.apache.org/jira/browse/HDFS-12482?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lei (Eddy) Xu updated HDFS-12482: - Attachment: HDFS-12482.04.patch Thanks for the suggestions, [~xiaochen] * Fixed checkstyle warnings. * Changed to use {{Preconditions}} for the value checks. * Ran the EC related failed tests above, all of them passed on my local machine. Will have an eye on the next jenkins run. > Provide a configuration to adjust the weight of EC recovery tasks to adjust > the speed of recovery > - > > Key: HDFS-12482 > URL: https://issues.apache.org/jira/browse/HDFS-12482 > Project: Hadoop HDFS > Issue Type: Improvement > Components: erasure-coding >Affects Versions: 3.0.0-alpha4 >Reporter: Lei (Eddy) Xu >Assignee: Lei (Eddy) Xu >Priority: Minor > Labels: hdfs-ec-3.0-nice-to-have > Attachments: HDFS-12482.00.patch, HDFS-12482.01.patch, > HDFS-12482.02.patch, HDFS-12482.03.patch, HDFS-12482.04.patch > > > The relative speed of EC recovery comparing to 3x replica recovery is a > function of (EC codec, number of sources, NIC speed, and CPU speed, and etc). > Currently the EC recovery has a fixed {{xmitsInProgress}} of {{max(# of > sources, # of targets)}} comparing to {{1}} for 3x replica recovery, and NN > uses {{xmitsInProgress}} to decide how much recovery tasks to schedule to the > DataNode this we can add a coefficient for user to tune the weight of EC > recovery tasks. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-12482) Provide a configuration to adjust the weight of EC recovery tasks to adjust the speed of recovery
[ https://issues.apache.org/jira/browse/HDFS-12482?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lei (Eddy) Xu updated HDFS-12482: - Attachment: HDFS-12482.03.patch thanks for the suggestions, [~xiaochen] Add checks as you suggested. > Provide a configuration to adjust the weight of EC recovery tasks to adjust > the speed of recovery > - > > Key: HDFS-12482 > URL: https://issues.apache.org/jira/browse/HDFS-12482 > Project: Hadoop HDFS > Issue Type: Improvement > Components: erasure-coding >Affects Versions: 3.0.0-alpha4 >Reporter: Lei (Eddy) Xu >Assignee: Lei (Eddy) Xu >Priority: Minor > Labels: hdfs-ec-3.0-nice-to-have > Attachments: HDFS-12482.00.patch, HDFS-12482.01.patch, > HDFS-12482.02.patch, HDFS-12482.03.patch > > > The relative speed of EC recovery comparing to 3x replica recovery is a > function of (EC codec, number of sources, NIC speed, and CPU speed, and etc). > Currently the EC recovery has a fixed {{xmitsInProgress}} of {{max(# of > sources, # of targets)}} comparing to {{1}} for 3x replica recovery, and NN > uses {{xmitsInProgress}} to decide how much recovery tasks to schedule to the > DataNode this we can add a coefficient for user to tune the weight of EC > recovery tasks. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-12482) Provide a configuration to adjust the weight of EC recovery tasks to adjust the speed of recovery
[ https://issues.apache.org/jira/browse/HDFS-12482?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lei (Eddy) Xu updated HDFS-12482: - Attachment: HDFS-12482.02.patch Thanks for the reviews, [~xiaochen] bq. suggest to add an example done bq. bump the timeout to at least 3 minutes bq. let's add input validation for xmitWeight. {{Math.max((int)(task.getXmits() * xmitWeight), 1);}} is very resilient to the input value. If the value is {{<= 0}}, it disables the weight, if it the value is larger enough, it basically only allow 1 task per period. Would it work for you? > Provide a configuration to adjust the weight of EC recovery tasks to adjust > the speed of recovery > - > > Key: HDFS-12482 > URL: https://issues.apache.org/jira/browse/HDFS-12482 > Project: Hadoop HDFS > Issue Type: Improvement > Components: erasure-coding >Affects Versions: 3.0.0-alpha4 >Reporter: Lei (Eddy) Xu >Assignee: Lei (Eddy) Xu >Priority: Minor > Labels: hdfs-ec-3.0-nice-to-have > Attachments: HDFS-12482.00.patch, HDFS-12482.01.patch, > HDFS-12482.02.patch > > > The relative speed of EC recovery comparing to 3x replica recovery is a > function of (EC codec, number of sources, NIC speed, and CPU speed, and etc). > Currently the EC recovery has a fixed {{xmitsInProgress}} of {{max(# of > sources, # of targets)}} comparing to {{1}} for 3x replica recovery, and NN > uses {{xmitsInProgress}} to decide how much recovery tasks to schedule to the > DataNode this we can add a coefficient for user to tune the weight of EC > recovery tasks. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-12482) Provide a configuration to adjust the weight of EC recovery tasks to adjust the speed of recovery
[ https://issues.apache.org/jira/browse/HDFS-12482?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lei (Eddy) Xu updated HDFS-12482: - Attachment: HDFS-12482.01.patch Thanks for the reviews, [~andrew.wang] Updated the patch to add documents. Empirically, an value between {{(0, 1.0]}} seems can achieve similar recovery speed and network / cpu overhead between replicated blocks and ec block recovery _on my testing cluster_. But it highly depends on HW. I will set this value to {{0.5}} initially in this patch. > Provide a configuration to adjust the weight of EC recovery tasks to adjust > the speed of recovery > - > > Key: HDFS-12482 > URL: https://issues.apache.org/jira/browse/HDFS-12482 > Project: Hadoop HDFS > Issue Type: Improvement > Components: erasure-coding >Affects Versions: 3.0.0-alpha4 >Reporter: Lei (Eddy) Xu >Assignee: Lei (Eddy) Xu >Priority: Minor > Labels: hdfs-ec-3.0-nice-to-have > Attachments: HDFS-12482.00.patch, HDFS-12482.01.patch > > > The relative speed of EC recovery comparing to 3x replica recovery is a > function of (EC codec, number of sources, NIC speed, and CPU speed, and etc). > Currently the EC recovery has a fixed {{xmitsInProgress}} of {{max(# of > sources, # of targets)}} comparing to {{1}} for 3x replica recovery, and NN > uses {{xmitsInProgress}} to decide how much recovery tasks to schedule to the > DataNode this we can add a coefficient for user to tune the weight of EC > recovery tasks. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-12482) Provide a configuration to adjust the weight of EC recovery tasks to adjust the speed of recovery
[ https://issues.apache.org/jira/browse/HDFS-12482?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lei (Eddy) Xu updated HDFS-12482: - Attachment: HDFS-12482.00.patch Add a weight to adjust xmits for performance. > Provide a configuration to adjust the weight of EC recovery tasks to adjust > the speed of recovery > - > > Key: HDFS-12482 > URL: https://issues.apache.org/jira/browse/HDFS-12482 > Project: Hadoop HDFS > Issue Type: Improvement > Components: erasure-coding >Affects Versions: 3.0.0-alpha4 >Reporter: Lei (Eddy) Xu >Assignee: Lei (Eddy) Xu >Priority: Minor > Labels: hdfs-ec-3.0-nice-to-have > Attachments: HDFS-12482.00.patch > > > The relative speed of EC recovery comparing to 3x replica recovery is a > function of (EC codec, number of sources, NIC speed, and CPU speed, and etc). > Currently the EC recovery has a fixed {{xmitsInProgress}} of {{max(# of > sources, # of targets)}} comparing to {{1}} for 3x replica recovery, and NN > uses {{xmitsInProgress}} to decide how much recovery tasks to schedule to the > DataNode this we can add a coefficient for user to tune the weight of EC > recovery tasks. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-12482) Provide a configuration to adjust the weight of EC recovery tasks to adjust the speed of recovery
[ https://issues.apache.org/jira/browse/HDFS-12482?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Lei (Eddy) Xu updated HDFS-12482: - Status: Patch Available (was: Open) > Provide a configuration to adjust the weight of EC recovery tasks to adjust > the speed of recovery > - > > Key: HDFS-12482 > URL: https://issues.apache.org/jira/browse/HDFS-12482 > Project: Hadoop HDFS > Issue Type: Improvement > Components: erasure-coding >Affects Versions: 3.0.0-alpha4 >Reporter: Lei (Eddy) Xu >Assignee: Lei (Eddy) Xu >Priority: Minor > Labels: hdfs-ec-3.0-nice-to-have > Attachments: HDFS-12482.00.patch > > > The relative speed of EC recovery comparing to 3x replica recovery is a > function of (EC codec, number of sources, NIC speed, and CPU speed, and etc). > Currently the EC recovery has a fixed {{xmitsInProgress}} of {{max(# of > sources, # of targets)}} comparing to {{1}} for 3x replica recovery, and NN > uses {{xmitsInProgress}} to decide how much recovery tasks to schedule to the > DataNode this we can add a coefficient for user to tune the weight of EC > recovery tasks. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
[jira] [Updated] (HDFS-12482) Provide a configuration to adjust the weight of EC recovery tasks to adjust the speed of recovery
[ https://issues.apache.org/jira/browse/HDFS-12482?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Andrew Wang updated HDFS-12482: --- Labels: hdfs-ec-3.0-nice-to-have (was: ) > Provide a configuration to adjust the weight of EC recovery tasks to adjust > the speed of recovery > - > > Key: HDFS-12482 > URL: https://issues.apache.org/jira/browse/HDFS-12482 > Project: Hadoop HDFS > Issue Type: Improvement > Components: erasure-coding >Affects Versions: 3.0.0-alpha4 >Reporter: Lei (Eddy) Xu >Assignee: Lei (Eddy) Xu >Priority: Minor > Labels: hdfs-ec-3.0-nice-to-have > > The relative speed of EC recovery comparing to 3x replica recovery is a > function of (EC codec, number of sources, NIC speed, and CPU speed, and etc). > Currently the EC recovery has a fixed {{xmitsInProgress}} of {{max(# of > sources, # of targets)}} comparing to {{1}} for 3x replica recovery, and NN > uses {{xmitsInProgress}} to decide how much recovery tasks to schedule to the > DataNode this we can add a coefficient for user to tune the weight of EC > recovery tasks. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org