[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13222337#comment-13222337 ] Hudson commented on MAPREDUCE-3897: --- Integrated in Hadoop-Mapreduce-trunk #1010 (See [https://builds.apache.org/job/Hadoop-Mapreduce-trunk/1010/]) MAPREDUCE-3897. Fixed computation of maxActiveAppsPerUser for queues by using capacity and not max-capacity since we are already scaling it by userLimitFactor. Contributed by Eric Payne. (Revision 1296898) Result = FAILURE acmurthy : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1296898 Files : * /hadoop/common/trunk/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestApplicationLimits.java > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Fix For: 0.23.2 > > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt, > MAPREDUCE-3897-2.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13222331#comment-13222331 ] Hudson commented on MAPREDUCE-3897: --- Integrated in Hadoop-Mapreduce-0.23-Build #216 (See [https://builds.apache.org/job/Hadoop-Mapreduce-0.23-Build/216/]) Merge -c 1296898 from trunk to branch-0.23 to fix MAPREDUCE-3897. Fixed computation of maxActiveAppsPerUser for queues by using capacity and not max-capacity since we are already scaling it by userLimitFactor. (Revision 1296899) Result = FAILURE acmurthy : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1296899 Files : * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestApplicationLimits.java > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Fix For: 0.23.2 > > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt, > MAPREDUCE-3897-2.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13222318#comment-13222318 ] Hudson commented on MAPREDUCE-3897: --- Integrated in Hadoop-Hdfs-0.23-Build #188 (See [https://builds.apache.org/job/Hadoop-Hdfs-0.23-Build/188/]) Merge -c 1296898 from trunk to branch-0.23 to fix MAPREDUCE-3897. Fixed computation of maxActiveAppsPerUser for queues by using capacity and not max-capacity since we are already scaling it by userLimitFactor. (Revision 1296899) Result = SUCCESS acmurthy : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1296899 Files : * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestApplicationLimits.java > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Fix For: 0.23.2 > > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt, > MAPREDUCE-3897-2.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13222312#comment-13222312 ] Hudson commented on MAPREDUCE-3897: --- Integrated in Hadoop-Hdfs-trunk #975 (See [https://builds.apache.org/job/Hadoop-Hdfs-trunk/975/]) MAPREDUCE-3897. Fixed computation of maxActiveAppsPerUser for queues by using capacity and not max-capacity since we are already scaling it by userLimitFactor. Contributed by Eric Payne. (Revision 1296898) Result = SUCCESS acmurthy : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1296898 Files : * /hadoop/common/trunk/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestApplicationLimits.java > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Fix For: 0.23.2 > > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt, > MAPREDUCE-3897-2.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13222087#comment-13222087 ] Hudson commented on MAPREDUCE-3897: --- Integrated in Hadoop-Mapreduce-trunk-Commit #1838 (See [https://builds.apache.org/job/Hadoop-Mapreduce-trunk-Commit/1838/]) MAPREDUCE-3897. Fixed computation of maxActiveAppsPerUser for queues by using capacity and not max-capacity since we are already scaling it by userLimitFactor. Contributed by Eric Payne. (Revision 1296898) Result = ABORTED acmurthy : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1296898 Files : * /hadoop/common/trunk/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestApplicationLimits.java > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Fix For: 0.23.2 > > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt, > MAPREDUCE-3897-2.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13222081#comment-13222081 ] Hudson commented on MAPREDUCE-3897: --- Integrated in Hadoop-Mapreduce-0.23-Commit #636 (See [https://builds.apache.org/job/Hadoop-Mapreduce-0.23-Commit/636/]) Merge -c 1296898 from trunk to branch-0.23 to fix MAPREDUCE-3897. Fixed computation of maxActiveAppsPerUser for queues by using capacity and not max-capacity since we are already scaling it by userLimitFactor. (Revision 1296899) Result = FAILURE acmurthy : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1296899 Files : * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestApplicationLimits.java > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Fix For: 0.23.2 > > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt, > MAPREDUCE-3897-2.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13222076#comment-13222076 ] Hudson commented on MAPREDUCE-3897: --- Integrated in Hadoop-Common-0.23-Commit #637 (See [https://builds.apache.org/job/Hadoop-Common-0.23-Commit/637/]) Merge -c 1296898 from trunk to branch-0.23 to fix MAPREDUCE-3897. Fixed computation of maxActiveAppsPerUser for queues by using capacity and not max-capacity since we are already scaling it by userLimitFactor. (Revision 1296899) Result = SUCCESS acmurthy : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1296899 Files : * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestApplicationLimits.java > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Fix For: 0.23.2 > > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt, > MAPREDUCE-3897-2.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13222073#comment-13222073 ] Hudson commented on MAPREDUCE-3897: --- Integrated in Hadoop-Common-trunk-Commit #1831 (See [https://builds.apache.org/job/Hadoop-Common-trunk-Commit/1831/]) MAPREDUCE-3897. Fixed computation of maxActiveAppsPerUser for queues by using capacity and not max-capacity since we are already scaling it by userLimitFactor. Contributed by Eric Payne. (Revision 1296898) Result = SUCCESS acmurthy : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1296898 Files : * /hadoop/common/trunk/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestApplicationLimits.java > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Fix For: 0.23.2 > > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt, > MAPREDUCE-3897-2.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13222069#comment-13222069 ] Hudson commented on MAPREDUCE-3897: --- Integrated in Hadoop-Hdfs-0.23-Commit #626 (See [https://builds.apache.org/job/Hadoop-Hdfs-0.23-Commit/626/]) Merge -c 1296898 from trunk to branch-0.23 to fix MAPREDUCE-3897. Fixed computation of maxActiveAppsPerUser for queues by using capacity and not max-capacity since we are already scaling it by userLimitFactor. (Revision 1296899) Result = SUCCESS acmurthy : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1296899 Files : * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java * /hadoop/common/branches/branch-0.23/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestApplicationLimits.java > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Fix For: 0.23.2 > > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt, > MAPREDUCE-3897-2.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13222071#comment-13222071 ] Hudson commented on MAPREDUCE-3897: --- Integrated in Hadoop-Hdfs-trunk-Commit #1905 (See [https://builds.apache.org/job/Hadoop-Hdfs-trunk-Commit/1905/]) MAPREDUCE-3897. Fixed computation of maxActiveAppsPerUser for queues by using capacity and not max-capacity since we are already scaling it by userLimitFactor. Contributed by Eric Payne. (Revision 1296898) Result = SUCCESS acmurthy : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1296898 Files : * /hadoop/common/trunk/hadoop-mapreduce-project/CHANGES.txt * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java * /hadoop/common/trunk/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestApplicationLimits.java > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Fix For: 0.23.2 > > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt, > MAPREDUCE-3897-2.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13221302#comment-13221302 ] Thomas Graves commented on MAPREDUCE-3897: -- +1 for the formula change to fix the current issue described here. I haven't had a chance to review the code in detail. I think there is still a bug in the computation in that it uses the user minimum limit % in the calculation of maxActiveApplicationPerUser, when really that need to factor in the number active users. so similar to computeUserLimit function where it takes the max of userlimit*capacity or capacity/activeusers. Perhaps Arun can comment. > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt, > MAPREDUCE-3897-2.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13221303#comment-13221303 ] Thomas Graves commented on MAPREDUCE-3897: -- I should add, my opinion is the other issue with user minimum limit % can be fixed in a separate jira. > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt, > MAPREDUCE-3897-2.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13221226#comment-13221226 ] Hadoop QA commented on MAPREDUCE-3897: -- +1 overall. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12516874/MAPREDUCE-3897-2.txt against trunk revision . +1 @author. The patch does not contain any @author tags. +1 tests included. The patch appears to include 3 new or modified tests. +1 javadoc. The javadoc tool did not generate any warning messages. +1 javac. The applied patch does not increase the total number of javac compiler warnings. +1 eclipse:eclipse. The patch built with eclipse:eclipse. +1 findbugs. The patch does not introduce any new Findbugs (version 1.3.9) warnings. +1 release audit. The applied patch does not increase the total number of release audit warnings. +1 core tests. The patch passed unit tests in . +1 contrib tests. The patch passed contrib unit tests. Test results: https://builds.apache.org/job/PreCommit-MAPREDUCE-Build/1986//testReport/ Console output: https://builds.apache.org/job/PreCommit-MAPREDUCE-Build/1986//console This message is automatically generated. > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt, > MAPREDUCE-3897-2.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13220240#comment-13220240 ] Thomas Graves commented on MAPREDUCE-3897: -- I just thought of a case where this won't work well for utilization. That is if you have a queue with small capacity - say 1%, but its max capacity is say 100%, even if we had the configuration per queue for am% and you set it really high, it might only be allowed a couple of AM's when in reality if the cluster has no one else running it should be allowed more so it could use the 100% max capacity. We might be better off leaving the maxActiveApplications computation using maxCapacity but changing the maxActiveApplicationsPerUser to use capacity and then allow the user limit factor to apply. Need to think about it some more. > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13219668#comment-13219668 ] Hadoop QA commented on MAPREDUCE-3897: -- +1 overall. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12516619/MAPREDUCE-3897-1.txt against trunk revision . +1 @author. The patch does not contain any @author tags. +1 tests included. The patch appears to include 3 new or modified tests. +1 javadoc. The javadoc tool did not generate any warning messages. +1 javac. The applied patch does not increase the total number of javac compiler warnings. +1 eclipse:eclipse. The patch built with eclipse:eclipse. +1 findbugs. The patch does not introduce any new Findbugs (version 1.3.9) warnings. +1 release audit. The applied patch does not increase the total number of release audit warnings. +1 core tests. The patch passed unit tests in . +1 contrib tests. The patch passed contrib unit tests. Test results: https://builds.apache.org/job/PreCommit-MAPREDUCE-Build/1973//testReport/ Console output: https://builds.apache.org/job/PreCommit-MAPREDUCE-Build/1973//console This message is automatically generated. > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > Attachments: MAPREDUCE-3897-1.txt, MAPREDUCE-3897-1.txt > > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (MAPREDUCE-3897) capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong
[ https://issues.apache.org/jira/browse/MAPREDUCE-3897?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13219523#comment-13219523 ] jirapos...@reviews.apache.org commented on MAPREDUCE-3897: -- --- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/4112/ --- Review request for Tom Graves. Summary --- MAPREDUCE-3897: capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong This addresses bug MAPREDUCE-3897. https://issues.apache.org/jira/browse/MAPREDUCE-3897 Diffs - branches/branch-0.23.2/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CapacitySchedulerConfiguration.java 1295253 branches/branch-0.23.2/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java 1295253 branches/branch-0.23.2/hadoop-mapreduce-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestApplicationLimits.java 1295253 Diff: https://reviews.apache.org/r/4112/diff Testing --- Thanks, Eric > capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong > -- > > Key: MAPREDUCE-3897 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-3897 > Project: Hadoop Map/Reduce > Issue Type: Bug > Components: mrv2 >Affects Versions: 0.23.0 >Reporter: Thomas Graves >Assignee: Eric Payne >Priority: Critical > > The capacity scheduler calculates the maxActiveApplications and the > maxActiveApplicationsPerUser based on the config > yarn.scheduler.capacity.maximum-applications or default 1. > MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * > maxAMResource% * absoluteMaxCapacity), 1) > MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * > (userLimit%/100) * userLimitFactor), 1) > maxActiveApplications is already multiplied by the queue absolute MAXIMUM > capacity, so if max capacity > capacity and if you have user limit factor 1 > (which is the default) and only 1 user is running, that user will not be > allowed to use over the queue capacity, so having it relative to MAX capacity > doesn't make sense. That user could easily end up in a deadlock and all its > space used by application masters. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira