[jira] [Commented] (YARN-4105) Capacity Scheduler headroom for DRF is wrong
[ https://issues.apache.org/jira/browse/YARN-4105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14731273#comment-14731273 ] Hudson commented on YARN-4105: -- FAILURE: Integrated in Hadoop-Hdfs-trunk #2273 (See [https://builds.apache.org/job/Hadoop-Hdfs-trunk/2273/]) YARN-4105. Capacity Scheduler headroom for DRF is wrong. Contributed by Chang Li (jlowe: rev 6eaca2e3634a88dc55689e8960352d6248c424d9) * hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestCapacityScheduler.java * hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java * hadoop-yarn-project/CHANGES.txt > Capacity Scheduler headroom for DRF is wrong > > > Key: YARN-4105 > URL: https://issues.apache.org/jira/browse/YARN-4105 > Project: Hadoop YARN > Issue Type: Bug > Components: capacityscheduler >Affects Versions: 2.6.0 >Reporter: Chang Li >Assignee: Chang Li > Fix For: 2.7.2 > > Attachments: YARN-4105.2.patch, YARN-4105.3.patch, YARN-4105.4.patch, > YARN-4105.patch > > > relate to the problem discussed in YARN-1857. But the min method is flawed > when we are using DRC. Have run into a real scenario in production where > queueCapacity: , qconsumed: vCores:361>, consumed: limit: vCores:755>. headRoom calculation returns 88064 where there is only 1536 > left in the queue because DRC effectively compare by vcores. It then caused > deadlock because RMcontainer allocator thought there is still space for > mapper and won't preempt a reducer in a full queue to schedule a mapper. > Propose fix with componentwiseMin. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (YARN-4105) Capacity Scheduler headroom for DRF is wrong
[ https://issues.apache.org/jira/browse/YARN-4105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14731234#comment-14731234 ] Hudson commented on YARN-4105: -- FAILURE: Integrated in Hadoop-Hdfs-trunk-Java8 #335 (See [https://builds.apache.org/job/Hadoop-Hdfs-trunk-Java8/335/]) YARN-4105. Capacity Scheduler headroom for DRF is wrong. Contributed by Chang Li (jlowe: rev 6eaca2e3634a88dc55689e8960352d6248c424d9) * hadoop-yarn-project/CHANGES.txt * hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java * hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestCapacityScheduler.java > Capacity Scheduler headroom for DRF is wrong > > > Key: YARN-4105 > URL: https://issues.apache.org/jira/browse/YARN-4105 > Project: Hadoop YARN > Issue Type: Bug > Components: capacityscheduler >Affects Versions: 2.6.0 >Reporter: Chang Li >Assignee: Chang Li > Fix For: 2.7.2 > > Attachments: YARN-4105.2.patch, YARN-4105.3.patch, YARN-4105.4.patch, > YARN-4105.patch > > > relate to the problem discussed in YARN-1857. But the min method is flawed > when we are using DRC. Have run into a real scenario in production where > queueCapacity: , qconsumed: vCores:361>, consumed: limit: vCores:755>. headRoom calculation returns 88064 where there is only 1536 > left in the queue because DRC effectively compare by vcores. It then caused > deadlock because RMcontainer allocator thought there is still space for > mapper and won't preempt a reducer in a full queue to schedule a mapper. > Propose fix with componentwiseMin. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (YARN-4105) Capacity Scheduler headroom for DRF is wrong
[ https://issues.apache.org/jira/browse/YARN-4105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14731203#comment-14731203 ] Hudson commented on YARN-4105: -- SUCCESS: Integrated in Hadoop-Mapreduce-trunk-Java8 #346 (See [https://builds.apache.org/job/Hadoop-Mapreduce-trunk-Java8/346/]) YARN-4105. Capacity Scheduler headroom for DRF is wrong. Contributed by Chang Li (jlowe: rev 6eaca2e3634a88dc55689e8960352d6248c424d9) * hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestCapacityScheduler.java * hadoop-yarn-project/CHANGES.txt * hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java > Capacity Scheduler headroom for DRF is wrong > > > Key: YARN-4105 > URL: https://issues.apache.org/jira/browse/YARN-4105 > Project: Hadoop YARN > Issue Type: Bug > Components: capacityscheduler >Affects Versions: 2.6.0 >Reporter: Chang Li >Assignee: Chang Li > Fix For: 2.7.2 > > Attachments: YARN-4105.2.patch, YARN-4105.3.patch, YARN-4105.4.patch, > YARN-4105.patch > > > relate to the problem discussed in YARN-1857. But the min method is flawed > when we are using DRC. Have run into a real scenario in production where > queueCapacity: , qconsumed: vCores:361>, consumed: limit: vCores:755>. headRoom calculation returns 88064 where there is only 1536 > left in the queue because DRC effectively compare by vcores. It then caused > deadlock because RMcontainer allocator thought there is still space for > mapper and won't preempt a reducer in a full queue to schedule a mapper. > Propose fix with componentwiseMin. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (YARN-4105) Capacity Scheduler headroom for DRF is wrong
[ https://issues.apache.org/jira/browse/YARN-4105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14731157#comment-14731157 ] Hudson commented on YARN-4105: -- FAILURE: Integrated in Hadoop-Mapreduce-trunk #2295 (See [https://builds.apache.org/job/Hadoop-Mapreduce-trunk/2295/]) YARN-4105. Capacity Scheduler headroom for DRF is wrong. Contributed by Chang Li (jlowe: rev 6eaca2e3634a88dc55689e8960352d6248c424d9) * hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestCapacityScheduler.java * hadoop-yarn-project/CHANGES.txt * hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java > Capacity Scheduler headroom for DRF is wrong > > > Key: YARN-4105 > URL: https://issues.apache.org/jira/browse/YARN-4105 > Project: Hadoop YARN > Issue Type: Bug > Components: capacityscheduler >Affects Versions: 2.6.0 >Reporter: Chang Li >Assignee: Chang Li > Fix For: 2.7.2 > > Attachments: YARN-4105.2.patch, YARN-4105.3.patch, YARN-4105.4.patch, > YARN-4105.patch > > > relate to the problem discussed in YARN-1857. But the min method is flawed > when we are using DRC. Have run into a real scenario in production where > queueCapacity: , qconsumed: vCores:361>, consumed: limit: vCores:755>. headRoom calculation returns 88064 where there is only 1536 > left in the queue because DRC effectively compare by vcores. It then caused > deadlock because RMcontainer allocator thought there is still space for > mapper and won't preempt a reducer in a full queue to schedule a mapper. > Propose fix with componentwiseMin. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (YARN-4105) Capacity Scheduler headroom for DRF is wrong
[ https://issues.apache.org/jira/browse/YARN-4105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14731103#comment-14731103 ] Hudson commented on YARN-4105: -- FAILURE: Integrated in Hadoop-Yarn-trunk #1083 (See [https://builds.apache.org/job/Hadoop-Yarn-trunk/1083/]) YARN-4105. Capacity Scheduler headroom for DRF is wrong. Contributed by Chang Li (jlowe: rev 6eaca2e3634a88dc55689e8960352d6248c424d9) * hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestCapacityScheduler.java * hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java * hadoop-yarn-project/CHANGES.txt > Capacity Scheduler headroom for DRF is wrong > > > Key: YARN-4105 > URL: https://issues.apache.org/jira/browse/YARN-4105 > Project: Hadoop YARN > Issue Type: Bug > Components: capacityscheduler >Affects Versions: 2.6.0 >Reporter: Chang Li >Assignee: Chang Li > Fix For: 2.7.2 > > Attachments: YARN-4105.2.patch, YARN-4105.3.patch, YARN-4105.4.patch, > YARN-4105.patch > > > relate to the problem discussed in YARN-1857. But the min method is flawed > when we are using DRC. Have run into a real scenario in production where > queueCapacity: , qconsumed: vCores:361>, consumed: limit: vCores:755>. headRoom calculation returns 88064 where there is only 1536 > left in the queue because DRC effectively compare by vcores. It then caused > deadlock because RMcontainer allocator thought there is still space for > mapper and won't preempt a reducer in a full queue to schedule a mapper. > Propose fix with componentwiseMin. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (YARN-4105) Capacity Scheduler headroom for DRF is wrong
[ https://issues.apache.org/jira/browse/YARN-4105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14730978#comment-14730978 ] Hudson commented on YARN-4105: -- FAILURE: Integrated in Hadoop-Yarn-trunk-Java8 #352 (See [https://builds.apache.org/job/Hadoop-Yarn-trunk-Java8/352/]) YARN-4105. Capacity Scheduler headroom for DRF is wrong. Contributed by Chang Li (jlowe: rev 6eaca2e3634a88dc55689e8960352d6248c424d9) * hadoop-yarn-project/CHANGES.txt * hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestCapacityScheduler.java * hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java > Capacity Scheduler headroom for DRF is wrong > > > Key: YARN-4105 > URL: https://issues.apache.org/jira/browse/YARN-4105 > Project: Hadoop YARN > Issue Type: Bug > Components: capacityscheduler >Affects Versions: 2.6.0 >Reporter: Chang Li >Assignee: Chang Li > Fix For: 2.7.2 > > Attachments: YARN-4105.2.patch, YARN-4105.3.patch, YARN-4105.4.patch, > YARN-4105.patch > > > relate to the problem discussed in YARN-1857. But the min method is flawed > when we are using DRC. Have run into a real scenario in production where > queueCapacity: , qconsumed: vCores:361>, consumed: limit: vCores:755>. headRoom calculation returns 88064 where there is only 1536 > left in the queue because DRC effectively compare by vcores. It then caused > deadlock because RMcontainer allocator thought there is still space for > mapper and won't preempt a reducer in a full queue to schedule a mapper. > Propose fix with componentwiseMin. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (YARN-4105) Capacity Scheduler headroom for DRF is wrong
[ https://issues.apache.org/jira/browse/YARN-4105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14730966#comment-14730966 ] Hudson commented on YARN-4105: -- FAILURE: Integrated in Hadoop-trunk-Commit #8403 (See [https://builds.apache.org/job/Hadoop-trunk-Commit/8403/]) YARN-4105. Capacity Scheduler headroom for DRF is wrong. Contributed by Chang Li (jlowe: rev 6eaca2e3634a88dc55689e8960352d6248c424d9) * hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java * hadoop-yarn-project/CHANGES.txt * hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestCapacityScheduler.java > Capacity Scheduler headroom for DRF is wrong > > > Key: YARN-4105 > URL: https://issues.apache.org/jira/browse/YARN-4105 > Project: Hadoop YARN > Issue Type: Bug > Components: capacityscheduler >Affects Versions: 2.6.0 >Reporter: Chang Li >Assignee: Chang Li > Fix For: 2.7.2 > > Attachments: YARN-4105.2.patch, YARN-4105.3.patch, YARN-4105.4.patch, > YARN-4105.patch > > > relate to the problem discussed in YARN-1857. But the min method is flawed > when we are using DRC. Have run into a real scenario in production where > queueCapacity: , qconsumed: vCores:361>, consumed: limit: vCores:755>. headRoom calculation returns 88064 where there is only 1536 > left in the queue because DRC effectively compare by vcores. It then caused > deadlock because RMcontainer allocator thought there is still space for > mapper and won't preempt a reducer in a full queue to schedule a mapper. > Propose fix with componentwiseMin. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (YARN-4105) Capacity Scheduler headroom for DRF is wrong
[ https://issues.apache.org/jira/browse/YARN-4105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14730906#comment-14730906 ] Jason Lowe commented on YARN-4105: -- Test failures are unrelated. Committing this. > Capacity Scheduler headroom for DRF is wrong > > > Key: YARN-4105 > URL: https://issues.apache.org/jira/browse/YARN-4105 > Project: Hadoop YARN > Issue Type: Bug > Components: capacityscheduler >Affects Versions: 2.6.0 >Reporter: Chang Li >Assignee: Chang Li > Attachments: YARN-4105.2.patch, YARN-4105.3.patch, YARN-4105.4.patch, > YARN-4105.patch > > > relate to the problem discussed in YARN-1857. But the min method is flawed > when we are using DRC. Have run into a real scenario in production where > queueCapacity: , qconsumed: vCores:361>, consumed: limit: vCores:755>. headRoom calculation returns 88064 where there is only 1536 > left in the queue because DRC effectively compare by vcores. It then caused > deadlock because RMcontainer allocator thought there is still space for > mapper and won't preempt a reducer in a full queue to schedule a mapper. > Propose fix with componentwiseMin. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (YARN-4105) Capacity Scheduler headroom for DRF is wrong
[ https://issues.apache.org/jira/browse/YARN-4105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14728386#comment-14728386 ] Hadoop QA commented on YARN-4105: - \\ \\ | (x) *{color:red}-1 overall{color}* | \\ \\ || Vote || Subsystem || Runtime || Comment || | {color:red}-1{color} | pre-patch | 17m 7s | Pre-patch trunk has 1 extant Findbugs (version 3.0.0) warnings. | | {color:green}+1{color} | @author | 0m 0s | The patch does not contain any @author tags. | | {color:green}+1{color} | tests included | 0m 0s | The patch appears to include 1 new or modified test files. | | {color:green}+1{color} | javac | 7m 50s | There were no new javac warning messages. | | {color:green}+1{color} | javadoc | 10m 3s | There were no new javadoc warning messages. | | {color:green}+1{color} | release audit | 0m 27s | The applied patch does not increase the total number of release audit warnings. | | {color:green}+1{color} | checkstyle | 0m 50s | There were no new checkstyle issues. | | {color:green}+1{color} | whitespace | 0m 1s | The patch has no lines that end in whitespace. | | {color:green}+1{color} | install | 1m 32s | mvn install still works. | | {color:green}+1{color} | eclipse:eclipse | 0m 34s | The patch built with eclipse:eclipse. | | {color:green}+1{color} | findbugs | 1m 32s | The patch does not introduce any new Findbugs (version 3.0.0) warnings. | | {color:red}-1{color} | yarn tests | 52m 20s | Tests failed in hadoop-yarn-server-resourcemanager. | | | | 92m 19s | | \\ \\ || Reason || Tests || | Failed unit tests | hadoop.yarn.server.resourcemanager.scheduler.fifo.TestFifoScheduler | | Timed out tests | org.apache.hadoop.yarn.server.resourcemanager.ahs.TestRMApplicationHistoryWriter | \\ \\ || Subsystem || Report/Notes || | Patch URL | http://issues.apache.org/jira/secure/attachment/12753876/YARN-4105.4.patch | | Optional Tests | javadoc javac unit findbugs checkstyle | | git revision | trunk / d31a41c | | Pre-patch Findbugs warnings | https://builds.apache.org/job/PreCommit-YARN-Build/8989/artifact/patchprocess/trunkFindbugsWarningshadoop-yarn-server-resourcemanager.html | | hadoop-yarn-server-resourcemanager test log | https://builds.apache.org/job/PreCommit-YARN-Build/8989/artifact/patchprocess/testrun_hadoop-yarn-server-resourcemanager.txt | | Test Results | https://builds.apache.org/job/PreCommit-YARN-Build/8989/testReport/ | | Java | 1.7.0_55 | | uname | Linux asf906.gq1.ygridcore.net 3.13.0-36-lowlatency #63-Ubuntu SMP PREEMPT Wed Sep 3 21:56:12 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux | | Console output | https://builds.apache.org/job/PreCommit-YARN-Build/8989/console | This message was automatically generated. > Capacity Scheduler headroom for DRF is wrong > > > Key: YARN-4105 > URL: https://issues.apache.org/jira/browse/YARN-4105 > Project: Hadoop YARN > Issue Type: Bug > Components: capacityscheduler >Affects Versions: 2.6.0 >Reporter: Chang Li >Assignee: Chang Li > Attachments: YARN-4105.2.patch, YARN-4105.3.patch, YARN-4105.4.patch, > YARN-4105.patch > > > relate to the problem discussed in YARN-1857. But the min method is flawed > when we are using DRC. Have run into a real scenario in production where > queueCapacity: , qconsumed: vCores:361>, consumed: limit: vCores:755>. headRoom calculation returns 88064 where there is only 1536 > left in the queue because DRC effectively compare by vcores. It then caused > deadlock because RMcontainer allocator thought there is still space for > mapper and won't preempt a reducer in a full queue to schedule a mapper. > Propose fix with componentwiseMin. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (YARN-4105) Capacity Scheduler headroom for DRF is wrong
[ https://issues.apache.org/jira/browse/YARN-4105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14728183#comment-14728183 ] Wangda Tan commented on YARN-4105: -- Patch LGTM too, thanks [~lichangleo]. Only one nit: could you update test comment to be: bq. // app 1 ask for 10GB memory and 1 vcore, To bq. // allocates 10GB memory and 1 vcore to app 1. Same to app2. > Capacity Scheduler headroom for DRF is wrong > > > Key: YARN-4105 > URL: https://issues.apache.org/jira/browse/YARN-4105 > Project: Hadoop YARN > Issue Type: Bug > Components: capacityscheduler >Affects Versions: 2.6.0 >Reporter: Chang Li >Assignee: Chang Li > Attachments: YARN-4105.2.patch, YARN-4105.3.patch, YARN-4105.patch > > > relate to the problem discussed in YARN-1857. But the min method is flawed > when we are using DRC. Have run into a real scenario in production where > queueCapacity: , qconsumed: vCores:361>, consumed: limit: vCores:755>. headRoom calculation returns 88064 where there is only 1536 > left in the queue because DRC effectively compare by vcores. It then caused > deadlock because RMcontainer allocator thought there is still space for > mapper and won't preempt a reducer in a full queue to schedule a mapper. > Propose fix with componentwiseMin. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (YARN-4105) Capacity Scheduler headroom for DRF is wrong
[ https://issues.apache.org/jira/browse/YARN-4105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14727758#comment-14727758 ] Hadoop QA commented on YARN-4105: - \\ \\ | (x) *{color:red}-1 overall{color}* | \\ \\ || Vote || Subsystem || Runtime || Comment || | {color:red}-1{color} | pre-patch | 17m 1s | Pre-patch trunk has 1 extant Findbugs (version 3.0.0) warnings. | | {color:green}+1{color} | @author | 0m 0s | The patch does not contain any @author tags. | | {color:green}+1{color} | tests included | 0m 0s | The patch appears to include 1 new or modified test files. | | {color:green}+1{color} | javac | 7m 53s | There were no new javac warning messages. | | {color:green}+1{color} | javadoc | 10m 7s | There were no new javadoc warning messages. | | {color:green}+1{color} | release audit | 0m 23s | The applied patch does not increase the total number of release audit warnings. | | {color:green}+1{color} | checkstyle | 0m 51s | There were no new checkstyle issues. | | {color:green}+1{color} | whitespace | 0m 0s | The patch has no lines that end in whitespace. | | {color:green}+1{color} | install | 1m 28s | mvn install still works. | | {color:green}+1{color} | eclipse:eclipse | 0m 33s | The patch built with eclipse:eclipse. | | {color:green}+1{color} | findbugs | 1m 27s | The patch does not introduce any new Findbugs (version 3.0.0) warnings. | | {color:red}-1{color} | yarn tests | 53m 55s | Tests failed in hadoop-yarn-server-resourcemanager. | | | | 93m 42s | | \\ \\ || Reason || Tests || | Failed unit tests | hadoop.yarn.server.resourcemanager.scheduler.fair.TestAllocationFileLoaderService | \\ \\ || Subsystem || Report/Notes || | Patch URL | http://issues.apache.org/jira/secure/attachment/12753789/YARN-4105.3.patch | | Optional Tests | javadoc javac unit findbugs checkstyle | | git revision | trunk / 7d6687f | | Pre-patch Findbugs warnings | https://builds.apache.org/job/PreCommit-YARN-Build/8980/artifact/patchprocess/trunkFindbugsWarningshadoop-yarn-server-resourcemanager.html | | hadoop-yarn-server-resourcemanager test log | https://builds.apache.org/job/PreCommit-YARN-Build/8980/artifact/patchprocess/testrun_hadoop-yarn-server-resourcemanager.txt | | Test Results | https://builds.apache.org/job/PreCommit-YARN-Build/8980/testReport/ | | Java | 1.7.0_55 | | uname | Linux asf905.gq1.ygridcore.net 3.13.0-36-lowlatency #63-Ubuntu SMP PREEMPT Wed Sep 3 21:56:12 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux | | Console output | https://builds.apache.org/job/PreCommit-YARN-Build/8980/console | This message was automatically generated. > Capacity Scheduler headroom for DRF is wrong > > > Key: YARN-4105 > URL: https://issues.apache.org/jira/browse/YARN-4105 > Project: Hadoop YARN > Issue Type: Bug > Components: capacityscheduler >Affects Versions: 2.6.0 >Reporter: Chang Li >Assignee: Chang Li > Attachments: YARN-4105.2.patch, YARN-4105.3.patch, YARN-4105.patch > > > relate to the problem discussed in YARN-1857. But the min method is flawed > when we are using DRC. Have run into a real scenario in production where > queueCapacity: , qconsumed: vCores:361>, consumed: limit: vCores:755>. headRoom calculation returns 88064 where there is only 1536 > left in the queue because DRC effectively compare by vcores. It then caused > deadlock because RMcontainer allocator thought there is still space for > mapper and won't preempt a reducer in a full queue to schedule a mapper. > Propose fix with componentwiseMin. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (YARN-4105) Capacity Scheduler headroom for DRF is wrong
[ https://issues.apache.org/jira/browse/YARN-4105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14727547#comment-14727547 ] Hadoop QA commented on YARN-4105: - \\ \\ | (x) *{color:red}-1 overall{color}* | \\ \\ || Vote || Subsystem || Runtime || Comment || | {color:red}-1{color} | pre-patch | 16m 23s | Pre-patch trunk has 1 extant Findbugs (version 3.0.0) warnings. | | {color:green}+1{color} | @author | 0m 0s | The patch does not contain any @author tags. | | {color:green}+1{color} | tests included | 0m 0s | The patch appears to include 1 new or modified test files. | | {color:green}+1{color} | javac | 7m 58s | There were no new javac warning messages. | | {color:green}+1{color} | javadoc | 10m 15s | There were no new javadoc warning messages. | | {color:green}+1{color} | release audit | 0m 23s | The applied patch does not increase the total number of release audit warnings. | | {color:green}+1{color} | checkstyle | 0m 52s | There were no new checkstyle issues. | | {color:red}-1{color} | whitespace | 0m 0s | The patch has 1 line(s) that end in whitespace. Use git apply --whitespace=fix. | | {color:green}+1{color} | install | 1m 29s | mvn install still works. | | {color:green}+1{color} | eclipse:eclipse | 0m 33s | The patch built with eclipse:eclipse. | | {color:green}+1{color} | findbugs | 1m 31s | The patch does not introduce any new Findbugs (version 3.0.0) warnings. | | {color:green}+1{color} | yarn tests | 53m 51s | Tests passed in hadoop-yarn-server-resourcemanager. | | | | 93m 19s | | \\ \\ || Subsystem || Report/Notes || | Patch URL | http://issues.apache.org/jira/secure/attachment/12753776/YARN-4105.2.patch | | Optional Tests | javadoc javac unit findbugs checkstyle | | git revision | trunk / 7d6687f | | Pre-patch Findbugs warnings | https://builds.apache.org/job/PreCommit-YARN-Build/8975/artifact/patchprocess/trunkFindbugsWarningshadoop-yarn-server-resourcemanager.html | | whitespace | https://builds.apache.org/job/PreCommit-YARN-Build/8975/artifact/patchprocess/whitespace.txt | | hadoop-yarn-server-resourcemanager test log | https://builds.apache.org/job/PreCommit-YARN-Build/8975/artifact/patchprocess/testrun_hadoop-yarn-server-resourcemanager.txt | | Test Results | https://builds.apache.org/job/PreCommit-YARN-Build/8975/testReport/ | | Java | 1.7.0_55 | | uname | Linux asf905.gq1.ygridcore.net 3.13.0-36-lowlatency #63-Ubuntu SMP PREEMPT Wed Sep 3 21:56:12 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux | | Console output | https://builds.apache.org/job/PreCommit-YARN-Build/8975/console | This message was automatically generated. > Capacity Scheduler headroom for DRF is wrong > > > Key: YARN-4105 > URL: https://issues.apache.org/jira/browse/YARN-4105 > Project: Hadoop YARN > Issue Type: Bug >Reporter: Chang Li >Assignee: Chang Li > Attachments: YARN-4105.2.patch, YARN-4105.patch > > > relate to the problem discussed in YARN-1857. But the min method is flawed > when we are using DRC. Have run into a real scenario in production where > queueCapacity: , qconsumed: vCores:361>, consumed: limit: vCores:755>. headRoom calculation returns 88064 where there is only 1536 > left in the queue because DRC effectively compare by vcores. It then caused > deadlock because RMcontainer allocator thought there is still space for > mapper and won't preempt a reducer in a full queue to schedule a mapper. > Propose fix with componentwiseMin. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (YARN-4105) Capacity Scheduler headroom for DRF is wrong
[ https://issues.apache.org/jira/browse/YARN-4105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14727427#comment-14727427 ] Chang Li commented on YARN-4105: the failed unit test is not related to my change. Have run the test on my machine with my patch on and the test passed. > Capacity Scheduler headroom for DRF is wrong > > > Key: YARN-4105 > URL: https://issues.apache.org/jira/browse/YARN-4105 > Project: Hadoop YARN > Issue Type: Bug >Reporter: Chang Li >Assignee: Chang Li > Attachments: YARN-4105.2.patch, YARN-4105.patch > > > relate to the problem discussed in YARN-1857. But the min method is flawed > when we are using DRC. Have run into a real scenario in production where > queueCapacity: , qconsumed: vCores:361>, consumed: limit: vCores:755>. headRoom calculation returns 88064 where there is only 1536 > left in the queue because DRC effectively compare by vcores. It then caused > deadlock because RMcontainer allocator thought there is still space for > mapper and won't preempt a reducer in a full queue to schedule a mapper. > Propose fix with componentwiseMin. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (YARN-4105) Capacity Scheduler headroom for DRF is wrong
[ https://issues.apache.org/jira/browse/YARN-4105?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14726782#comment-14726782 ] Hadoop QA commented on YARN-4105: - \\ \\ | (x) *{color:red}-1 overall{color}* | \\ \\ || Vote || Subsystem || Runtime || Comment || | {color:red}-1{color} | pre-patch | 17m 0s | Pre-patch trunk has 1 extant Findbugs (version 3.0.0) warnings. | | {color:green}+1{color} | @author | 0m 0s | The patch does not contain any @author tags. | | {color:green}+1{color} | tests included | 0m 0s | The patch appears to include 1 new or modified test files. | | {color:green}+1{color} | javac | 7m 47s | There were no new javac warning messages. | | {color:green}+1{color} | javadoc | 9m 58s | There were no new javadoc warning messages. | | {color:green}+1{color} | release audit | 0m 24s | The applied patch does not increase the total number of release audit warnings. | | {color:green}+1{color} | checkstyle | 0m 51s | There were no new checkstyle issues. | | {color:red}-1{color} | whitespace | 0m 0s | The patch has 10 line(s) that end in whitespace. Use git apply --whitespace=fix. | | {color:green}+1{color} | install | 1m 26s | mvn install still works. | | {color:green}+1{color} | eclipse:eclipse | 0m 33s | The patch built with eclipse:eclipse. | | {color:green}+1{color} | findbugs | 1m 29s | The patch does not introduce any new Findbugs (version 3.0.0) warnings. | | {color:red}-1{color} | yarn tests | 54m 17s | Tests failed in hadoop-yarn-server-resourcemanager. | | | | 93m 49s | | \\ \\ || Reason || Tests || | Failed unit tests | hadoop.yarn.server.resourcemanager.scheduler.fifo.TestFifoScheduler | \\ \\ || Subsystem || Report/Notes || | Patch URL | http://issues.apache.org/jira/secure/attachment/12753664/YARN-4105.patch | | Optional Tests | javadoc javac unit findbugs checkstyle | | git revision | trunk / 00804e2 | | Pre-patch Findbugs warnings | https://builds.apache.org/job/PreCommit-YARN-Build/8963/artifact/patchprocess/trunkFindbugsWarningshadoop-yarn-server-resourcemanager.html | | whitespace | https://builds.apache.org/job/PreCommit-YARN-Build/8963/artifact/patchprocess/whitespace.txt | | hadoop-yarn-server-resourcemanager test log | https://builds.apache.org/job/PreCommit-YARN-Build/8963/artifact/patchprocess/testrun_hadoop-yarn-server-resourcemanager.txt | | Test Results | https://builds.apache.org/job/PreCommit-YARN-Build/8963/testReport/ | | Java | 1.7.0_55 | | uname | Linux asf901.gq1.ygridcore.net 3.13.0-36-lowlatency #63-Ubuntu SMP PREEMPT Wed Sep 3 21:56:12 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux | | Console output | https://builds.apache.org/job/PreCommit-YARN-Build/8963/console | This message was automatically generated. > Capacity Scheduler headroom for DRF is wrong > > > Key: YARN-4105 > URL: https://issues.apache.org/jira/browse/YARN-4105 > Project: Hadoop YARN > Issue Type: Bug >Reporter: Chang Li >Assignee: Chang Li > Attachments: YARN-4105.patch > > > relate to the problem discussed in YARN-1857. But the min method is flawed > when we are using DRC. Have run into a real scenario in production where > queueCapacity: , qconsumed: vCores:361>, consumed: limit: vCores:755>. headRoom calculation returns 88064 where there is only 1536 > left in the queue because DRC effectively compare by vcores. It then caused > deadlock because RMcontainer allocator thought there is still space for > mapper and won't preempt a reducer in a full queue to schedule a mapper. > Propose fix with componentwiseMin. -- This message was sent by Atlassian JIRA (v6.3.4#6332)