[jira] [Commented] (YARN-6964) Fair scheduler misuses Resources operations

[ https://issues.apache.org/jira/browse/YARN-6964?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16149241#comment-16149241 ] Karthik Kambatla commented on YARN-6964: [~templedf] - can you quickly check if the test failures

[jira] [Assigned] (YARN-3870) Providing raw container request information for fine scheduling

[ https://issues.apache.org/jira/browse/YARN-3870?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla reassigned YARN-3870: -- Assignee: (was: Karthik Kambatla) > Providing raw container request information for

[jira] [Assigned] (YARN-4487) FairScheduler: Compute per-container allocation latency and roll up to get per-application and per-queue

[ https://issues.apache.org/jira/browse/YARN-4487?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla reassigned YARN-4487: -- Assignee: (was: Karthik Kambatla) > FairScheduler: Compute per-container allocation

[jira] [Assigned] (YARN-4486) Add requestTime to ResourceRequest

[ https://issues.apache.org/jira/browse/YARN-4486?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla reassigned YARN-4486: -- Assignee: (was: Karthik Kambatla) > Add requestTime to ResourceRequest >

[jira] [Commented] (YARN-4509) Promote containers from OPPORTUNISTIC to GUARANTEED

[ https://issues.apache.org/jira/browse/YARN-4509?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16149226#comment-16149226 ] Karthik Kambatla commented on YARN-4509: [~asuresh] - didn't you already add a way to do this? /cc

[jira] [Resolved] (YARN-2814) RM: Clean-up the handling of "fatal" events

[ https://issues.apache.org/jira/browse/YARN-2814?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla resolved YARN-2814. Resolution: Not A Problem I believe most of the items I called out had been fixed over

[jira] [Assigned] (YARN-2963) Helper library that allows requesting containers from multiple queues

[ https://issues.apache.org/jira/browse/YARN-2963?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla reassigned YARN-2963: -- Assignee: (was: Karthik Kambatla) > Helper library that allows requesting

[jira] [Commented] (YARN-7123) FairScheduler.getResourceCalculator() returns an instance of DefaultResourceCalculator regardless of the configuration

[ https://issues.apache.org/jira/browse/YARN-7123?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16149216#comment-16149216 ] Karthik Kambatla commented on YARN-7123: Are there cases where we should not consider all the

[jira] [Commented] (YARN-6964) Fair scheduler misuses Resources operations

[ https://issues.apache.org/jira/browse/YARN-6964?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16148212#comment-16148212 ] Karthik Kambatla commented on YARN-6964: Taking a fresh look. Only one comment: do we need

[jira] [Commented] (YARN-7057) FSAppAttempt#getResourceUsage doesn't need to consider resources queued for preemption

[ https://issues.apache.org/jira/browse/YARN-7057?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16142512#comment-16142512 ] Karthik Kambatla commented on YARN-7057: [~templedf] - good questions. Reasons why

[jira] [Commented] (YARN-7086) Release all containers aynchronously

[ https://issues.apache.org/jira/browse/YARN-7086?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16138947#comment-16138947 ] Karthik Kambatla commented on YARN-7086: I like Jason's idea. > Release all containers

[jira] [Commented] (YARN-6798) Fix NM startup failure with old state store due to version mismatch

[ https://issues.apache.org/jira/browse/YARN-6798?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16137593#comment-16137593 ] Karthik Kambatla commented on YARN-6798: Nit pick on the patch: the second line says {{1.2 to

[jira] [Commented] (YARN-6127) Add support for work preserving NM restart when AMRMProxy is enabled

[ https://issues.apache.org/jira/browse/YARN-6127?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16136099#comment-16136099 ] Karthik Kambatla commented on YARN-6127: Given this breaks rolling upgrade, are we sure we want to

[jira] [Updated] (YARN-7049) FSAppAttempt preemption related fields have confusing names

[ https://issues.apache.org/jira/browse/YARN-7049?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-7049: --- Attachment: YARN-7049.002.patch v2 patch fixes the checkstyle issue. No tests since no change

[jira] [Commented] (YARN-7057) FSAppAttempt#getResourceUsage doesn't need to consider resources queued for preemption

[ https://issues.apache.org/jira/browse/YARN-7057?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16134487#comment-16134487 ] Karthik Kambatla commented on YARN-7057: The test failure in CapacityScheduler should be unrelated.

[jira] [Updated] (YARN-7049) FSAppAttempt preemption related fields have confusing names

[ https://issues.apache.org/jira/browse/YARN-7049?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-7049: --- Attachment: YARN-7049.001.patch Trivial patch that renames the fields in question. >

[jira] [Updated] (YARN-7049) FSAppAttempt preemption related fields have confusing names

[ https://issues.apache.org/jira/browse/YARN-7049?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-7049: --- Target Version/s: 2.9.0, 3.0.0-beta1 (was: 3.0.0-beta1) > FSAppAttempt preemption related

[jira] [Updated] (YARN-7049) FSAppAttempt preemption related fields have confusing names

[ https://issues.apache.org/jira/browse/YARN-7049?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-7049: --- Summary: FSAppAttempt preemption related fields have confusing names (was: FSAppAttempt

[jira] [Updated] (YARN-7057) FSAppAttempt#getResourceUsage doesn't need to consider resources queued for preemption

[ https://issues.apache.org/jira/browse/YARN-7057?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-7057: --- Summary: FSAppAttempt#getResourceUsage doesn't need to consider resources queued for

[jira] [Updated] (YARN-7057) FSAppAttempt#getResourceUsage doesn't need to consider resources queued for preemption

[ https://issues.apache.org/jira/browse/YARN-7057?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-7057: --- Issue Type: Improvement (was: Bug) > FSAppAttempt#getResourceUsage doesn't need to consider

[jira] [Updated] (YARN-7057) FSAppAttempt#getResourceUsage doesn't need to considered resources queued for preemption

[ https://issues.apache.org/jira/browse/YARN-7057?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-7057: --- Attachment: YARN-7057.001.patch Patch simplifies getResourceUsage and updates

[jira] [Created] (YARN-7057) FSAppAttempt#getResourceUsage doesn't need to considered resources queued for preemption

Karthik Kambatla created YARN-7057: -- Summary: FSAppAttempt#getResourceUsage doesn't need to considered resources queued for preemption Key: YARN-7057 URL: https://issues.apache.org/jira/browse/YARN-7057

[jira] [Created] (YARN-7049) FSAppAttempt fields tracking containers/resources queued for preemption can use better names

Karthik Kambatla created YARN-7049: -- Summary: FSAppAttempt fields tracking containers/resources queued for preemption can use better names Key: YARN-7049 URL: https://issues.apache.org/jira/browse/YARN-7049

[jira] [Commented] (YARN-6964) Fair scheduler misuses Resources operations

[ https://issues.apache.org/jira/browse/YARN-6964?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16124354#comment-16124354 ] Karthik Kambatla commented on YARN-6964: Thanks for looking into these, Daniel. Comments on the

[jira] [Comment Edited] (YARN-6964) Fair scheduler misuses Resources operations

[ https://issues.apache.org/jira/browse/YARN-6964?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16124354#comment-16124354 ] Karthik Kambatla edited comment on YARN-6964 at 8/12/17 12:36 AM: -- Thanks

[jira] [Commented] (YARN-6956) preemption may only consider resource requests for one node

[ https://issues.apache.org/jira/browse/YARN-6956?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16124069#comment-16124069 ] Karthik Kambatla commented on YARN-6956: bq. If so, how do we avoid preempting small amounts at a

[jira] [Commented] (YARN-6760) ZKRMStateStore.constructZkRootNodeACL seems to not set acls as desired

[ https://issues.apache.org/jira/browse/YARN-6760?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16123882#comment-16123882 ] Karthik Kambatla commented on YARN-6760: Is there a particular ill-effect of said issue with ACLs?

[jira] [Updated] (YARN-6974) Make CuratorBasedElectorService the default

[ https://issues.apache.org/jira/browse/YARN-6974?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-6974: --- Priority: Critical (was: Major) > Make CuratorBasedElectorService the default >

[jira] [Commented] (YARN-6956) preemption may only consider resource requests for one node

[ https://issues.apache.org/jira/browse/YARN-6956?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16117772#comment-16117772 ] Karthik Kambatla commented on YARN-6956: There are a series of follow-up changes to be made: #

[jira] [Commented] (YARN-6705) Add separate NM preemption thresholds for cpu and memory

[ https://issues.apache.org/jira/browse/YARN-6705?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16084457#comment-16084457 ] Karthik Kambatla commented on YARN-6705: +1, pending Jenkins. > Add separate NM preemption

[jira] [Commented] (YARN-6705) Add separate NM preemption thresholds for cpu and memory

[ https://issues.apache.org/jira/browse/YARN-6705?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16084433#comment-16084433 ] Karthik Kambatla commented on YARN-6705: Thanks Haibo. Just one nit around config names: Can we

[jira] [Commented] (YARN-6706) Refactor ContainerScheduler to make oversubscription change easier

[ https://issues.apache.org/jira/browse/YARN-6706?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16080813#comment-16080813 ] Karthik Kambatla commented on YARN-6706: I feel we should get this into trunk directly instead of

[jira] [Commented] (YARN-6764) Simplify the logic in FairScheduler#attemptScheduling

[ https://issues.apache.org/jira/browse/YARN-6764?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16079732#comment-16079732 ] Karthik Kambatla commented on YARN-6764: +1 > Simplify the logic in

[jira] [Commented] (YARN-6670) Add separate NM overallocation thresholds for cpu and memory

[ https://issues.apache.org/jira/browse/YARN-6670?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16054487#comment-16054487 ] Karthik Kambatla commented on YARN-6670: I like the idea of making a decision based on the usage

[jira] [Comment Edited] (YARN-6670) Add separate NM overallocation thresholds for cpu and memory

[ https://issues.apache.org/jira/browse/YARN-6670?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16045041#comment-16045041 ] Karthik Kambatla edited comment on YARN-6670 at 6/9/17 9:19 PM: +1, pending

[jira] [Commented] (YARN-6670) Add separate NM overallocation thresholds for cpu and memory

[ https://issues.apache.org/jira/browse/YARN-6670?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16045041#comment-16045041 ] Karthik Kambatla commented on YARN-6670: +1, pending Jenkins. [~ kkaranasos] - do you want to take

[jira] [Commented] (YARN-6670) Add separate NM overallocation thresholds for cpu and memory

[ https://issues.apache.org/jira/browse/YARN-6670?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16044961#comment-16044961 ] Karthik Kambatla commented on YARN-6670: Thanks for updating the patch, Haibo. Looks mostly good.

[jira] [Commented] (YARN-6670) Add separate NM overallocation thresholds for cpu and memory

[ https://issues.apache.org/jira/browse/YARN-6670?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16035468#comment-16035468 ] Karthik Kambatla commented on YARN-6670: Couple of comments: # I think there is value in keeping

[jira] [Updated] (YARN-6246) Identifying starved apps does not need the scheduler writelock

[ https://issues.apache.org/jira/browse/YARN-6246?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-6246: --- Attachment: YARN-6246.005.patch Looked at it again with fresh eyes and the long-term safer

[jira] [Commented] (YARN-6246) Identifying starved apps does not need the scheduler writelock

[ https://issues.apache.org/jira/browse/YARN-6246?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16023954#comment-16023954 ] Karthik Kambatla commented on YARN-6246: Updated patch uses an AtomicLong for

[jira] [Updated] (YARN-6246) Identifying starved apps does not need the scheduler writelock

[ https://issues.apache.org/jira/browse/YARN-6246?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-6246: --- Attachment: YARN-6246.004.patch > Identifying starved apps does not need the scheduler

[jira] [Commented] (YARN-5531) UnmanagedAM pool manager for federating application across clusters

[ https://issues.apache.org/jira/browse/YARN-5531?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16023443#comment-16023443 ] Karthik Kambatla commented on YARN-5531: [~botong] - apologies for not getting to the review. I

[jira] [Commented] (YARN-5531) UnmanagedAM pool manager for federating application across clusters

[ https://issues.apache.org/jira/browse/YARN-5531?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16021664#comment-16021664 ] Karthik Kambatla commented on YARN-5531: Sorry for the delay on this, Botong. Will take a look

[jira] [Commented] (YARN-6602) Impersonation does not work if standby RM is contacted first

[ https://issues.apache.org/jira/browse/YARN-6602?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16016493#comment-16016493 ] Karthik Kambatla commented on YARN-6602: +1 on the latest patch. Let us wait for a day or two for

[jira] [Commented] (YARN-6602) Impersonation does not work if standby RM is contacted first

[ https://issues.apache.org/jira/browse/YARN-6602?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16014997#comment-16014997 ] Karthik Kambatla commented on YARN-6602: The patch looks good to me. One minor comment:

[jira] [Resolved] (YARN-6590) ResourceManager Master/Slave transition make all applications killed

[ https://issues.apache.org/jira/browse/YARN-6590?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla resolved YARN-6590. Resolution: Invalid [~wuchang1989], in the future, please reach out on our user mailing list

[jira] [Commented] (YARN-6582) FSAppAttempt demand can be updated atomically in updateDemand()

[ https://issues.apache.org/jira/browse/YARN-6582?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16006781#comment-16006781 ] Karthik Kambatla commented on YARN-6582: The patch doesn't change the behavior, hence no tests. >

[jira] [Updated] (YARN-6582) FSAppAttempt demand can be updated atomically in updateDemand()

[ https://issues.apache.org/jira/browse/YARN-6582?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-6582: --- Attachment: YARN-6582.001.patch Uploading patch to improve this: # Use a new variable

[jira] [Commented] (YARN-6246) Identifying starved apps does not need the scheduler writelock

[ https://issues.apache.org/jira/browse/YARN-6246?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16005727#comment-16005727 ] Karthik Kambatla commented on YARN-6246: Carefully looked through the locking. dumpStateInternal

[jira] [Created] (YARN-6582) FSAppAttempt demand can be updated atomically in updateDemand()

Karthik Kambatla created YARN-6582: -- Summary: FSAppAttempt demand can be updated atomically in updateDemand() Key: YARN-6582 URL: https://issues.apache.org/jira/browse/YARN-6582 Project: Hadoop YARN

[jira] [Commented] (YARN-6486) FairScheduler: Deprecate continuous scheduling in 2.9

[ https://issues.apache.org/jira/browse/YARN-6486?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16005704#comment-16005704 ] Karthik Kambatla commented on YARN-6486: The non-HB driven approach allows certain optimizations.

[jira] [Commented] (YARN-6380) FSAppAttempt keeps redundant copy of the queue

[ https://issues.apache.org/jira/browse/YARN-6380?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16005592#comment-16005592 ] Karthik Kambatla commented on YARN-6380: I am torn between following the style guide in spirit vs

[jira] [Commented] (YARN-5531) UnmanagedAM pool manager for federating application across clusters

[ https://issues.apache.org/jira/browse/YARN-5531?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16005569#comment-16005569 ] Karthik Kambatla commented on YARN-5531: Thanks for working on this, [~botong]. I took a close look

[jira] [Commented] (YARN-6380) FSAppAttempt keeps redundant copy of the queue

[ https://issues.apache.org/jira/browse/YARN-6380?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16003834#comment-16003834 ] Karthik Kambatla commented on YARN-6380: At least one of the checkstyle issues looks legit. Mind

[jira] [Commented] (YARN-6380) FSAppAttempt keeps redundant copy of the queue

[ https://issues.apache.org/jira/browse/YARN-6380?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16003641#comment-16003641 ] Karthik Kambatla commented on YARN-6380: +1, pending Jenkins. > FSAppAttempt keeps redundant copy

[jira] [Commented] (YARN-3742) YARN RM will shut down if ZKClient creation times out

[ https://issues.apache.org/jira/browse/YARN-3742?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16003625#comment-16003625 ] Karthik Kambatla commented on YARN-3742: Just committed to trunk and branch-2. Thanks for fixing

[jira] [Commented] (YARN-3742) YARN RM will shut down if ZKClient creation times out

[ https://issues.apache.org/jira/browse/YARN-3742?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16003614#comment-16003614 ] Karthik Kambatla commented on YARN-3742: Missed that. Looks good. +1. > YARN RM will shut down

[jira] [Commented] (YARN-1197) Support changing resources of an allocated container

[ https://issues.apache.org/jira/browse/YARN-1197?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16001656#comment-16001656 ] Karthik Kambatla commented on YARN-1197: FairScheduler supports strict locality (relaxLocality ==

[jira] [Commented] (YARN-3839) Quit throwing NMNotYetReadyException

[ https://issues.apache.org/jira/browse/YARN-3839?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16001468#comment-16001468 ] Karthik Kambatla commented on YARN-3839: Thanks [~maniraj...@gmail.com] for working on this, and

[jira] [Commented] (YARN-6565) Fix memory leak and finish app trigger in AMRMProxy

[ https://issues.apache.org/jira/browse/YARN-6565?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16001454#comment-16001454 ] Karthik Kambatla commented on YARN-6565: Bumped up priority cause memory leak does sound severe.

[jira] [Updated] (YARN-6565) Fix memory leak and finish app trigger in AMRMProxy

[ https://issues.apache.org/jira/browse/YARN-6565?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-6565: --- Priority: Critical (was: Minor) > Fix memory leak and finish app trigger in AMRMProxy >

[jira] [Commented] (YARN-3742) YARN RM will shut down if ZKClient creation times out

[ https://issues.apache.org/jira/browse/YARN-3742?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15979612#comment-15979612 ] Karthik Kambatla commented on YARN-3742: That looks much cleaner. Thanks for updating the patch.

[jira] [Commented] (YARN-6433) Only accessible cgroup mount directories should be selected for a controller

[ https://issues.apache.org/jira/browse/YARN-6433?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15969588#comment-15969588 ] Karthik Kambatla commented on YARN-6433: +1 > Only accessible cgroup mount directories should be

[jira] [Commented] (YARN-6432) FairScheduler: Reserve preempted resources for corresponding applications

[ https://issues.apache.org/jira/browse/YARN-6432?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15966706#comment-15966706 ] Karthik Kambatla commented on YARN-6432: Committed to trunk. branch-2 patch has compilation issues.

[jira] [Updated] (YARN-6432) FairScheduler: Reserve preempted resources for corresponding applications

[ https://issues.apache.org/jira/browse/YARN-6432?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-6432: --- Summary: FairScheduler: Reserve preempted resources for corresponding applications (was: FS

[jira] [Commented] (YARN-6432) FS preemption should reserve a node before considering containers on it for preemption

[ https://issues.apache.org/jira/browse/YARN-6432?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15966637#comment-15966637 ] Karthik Kambatla commented on YARN-6432: +1, checking this in. > FS preemption should reserve a

[jira] [Updated] (YARN-6298) Metric preemptCall is not used in new preemption

[ https://issues.apache.org/jira/browse/YARN-6298?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-6298: --- Summary: Metric preemptCall is not used in new preemption (was: Metric preemptCall is not

[jira] [Commented] (YARN-6298) Metric preemptCall is not used in new preemption.

[ https://issues.apache.org/jira/browse/YARN-6298?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15961541#comment-15961541 ] Karthik Kambatla commented on YARN-6298: Unfortunately, the patch doesn't apply cleanly anymore.

[jira] [Commented] (YARN-3742) YARN RM will shut down if ZKClient creation times out

[ https://issues.apache.org/jira/browse/YARN-3742?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15961537#comment-15961537 ] Karthik Kambatla commented on YARN-3742: I like the clarity in the log messages. However, the cases

[jira] [Commented] (YARN-6298) Metric preemptCall is not used in new preemption.

[ https://issues.apache.org/jira/browse/YARN-6298?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15961527#comment-15961527 ] Karthik Kambatla commented on YARN-6298: +1. Checking this in. > Metric preemptCall is not used

[jira] [Commented] (YARN-6433) Only accessible cgroup mount directories should be selected for a controller

[ https://issues.apache.org/jira/browse/YARN-6433?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15961524#comment-15961524 ] Karthik Kambatla commented on YARN-6433: Patch looks reasonable. Some minor comments: # File has a

[jira] [Resolved] (YARN-6448) Continuous scheduling thread crashes while sorting nodes

[ https://issues.apache.org/jira/browse/YARN-6448?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla resolved YARN-6448. Resolution: Fixed Hadoop Flags: Reviewed Fix Version/s: 2.9.0 Just committed

[jira] [Updated] (YARN-6448) Continuous scheduling thread crashes while sorting nodes

[ https://issues.apache.org/jira/browse/YARN-6448?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-6448: --- Summary: Continuous scheduling thread crashes while sorting nodes (was: Continuous scheduling

[jira] [Commented] (YARN-6448) Continuous scheduling thread crash while sorting nodes

[ https://issues.apache.org/jira/browse/YARN-6448?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15957939#comment-15957939 ] Karthik Kambatla commented on YARN-6448: +1. Committing this based on the Jenkins run against v2

[jira] [Commented] (YARN-6449) Enable YARN to accept jobs with < 1 core allocations

[ https://issues.apache.org/jira/browse/YARN-6449?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15957687#comment-15957687 ] Karthik Kambatla commented on YARN-6449: Which scheduler are you using? FairScheduler already

[jira] [Commented] (YARN-5531) UnmanagedAM pool manager for federating application across clusters

[ https://issues.apache.org/jira/browse/YARN-5531?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15956086#comment-15956086 ] Karthik Kambatla commented on YARN-5531: Barely skimmed through the patch. Some high-level

[jira] [Commented] (YARN-6436) TestSchedulingPolicy#testParseSchedulingPolicy timeout is too low

[ https://issues.apache.org/jira/browse/YARN-6436?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15955783#comment-15955783 ] Karthik Kambatla commented on YARN-6436: +1, checking this in. >

[jira] [Comment Edited] (YARN-6245) Add FinalResource object to reduce overhead of Resource class instancing

[ https://issues.apache.org/jira/browse/YARN-6245?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15955780#comment-15955780 ] Karthik Kambatla edited comment on YARN-6245 at 4/4/17 8:47 PM: bq. It

[jira] [Commented] (YARN-6245) Add FinalResource object to reduce overhead of Resource class instancing

[ https://issues.apache.org/jira/browse/YARN-6245?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15955780#comment-15955780 ] Karthik Kambatla commented on YARN-6245: bq. It works very well so I think we may not need a huge

[jira] [Updated] (YARN-5829) FS preemption should reserve a node before considering containers on it for preemption

[ https://issues.apache.org/jira/browse/YARN-5829?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-5829: --- Issue Type: Bug (was: Sub-task) Parent: (was: YARN-5990) > FS preemption should

[jira] [Updated] (YARN-6432) FS preemption should reserve a node before considering containers on it for preemption

[ https://issues.apache.org/jira/browse/YARN-6432?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-6432: --- Issue Type: Sub-task (was: Bug) Parent: YARN-5990 > FS preemption should reserve a

[jira] [Commented] (YARN-6245) Add FinalResource object to reduce overhead of Resource class instancing

[ https://issues.apache.org/jira/browse/YARN-6245?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15954477#comment-15954477 ] Karthik Kambatla commented on YARN-6245: One other thing. The scheduler should probably only use

[jira] [Commented] (YARN-6245) Add FinalResource object to reduce overhead of Resource class instancing

[ https://issues.apache.org/jira/browse/YARN-6245?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15954300#comment-15954300 ] Karthik Kambatla commented on YARN-6245: We are in agreement. {{LightResource}} seems very useful

[jira] [Commented] (YARN-6245) Add FinalResource object to reduce overhead of Resource class instancing

[ https://issues.apache.org/jira/browse/YARN-6245?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15954232#comment-15954232 ] Karthik Kambatla commented on YARN-6245: I think I understand some more now. :) Now, I see how the

[jira] [Commented] (YARN-6245) Add FinalResource object to reduce overhead of Resource class instancing

[ https://issues.apache.org/jira/browse/YARN-6245?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15954014#comment-15954014 ] Karthik Kambatla commented on YARN-6245: [~leftnoteasy] - I see the difference between

[jira] [Comment Edited] (YARN-6245) Add FinalResource object to reduce overhead of Resource class instancing

[ https://issues.apache.org/jira/browse/YARN-6245?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15954014#comment-15954014 ] Karthik Kambatla edited comment on YARN-6245 at 4/3/17 7:02 PM:

[jira] [Commented] (YARN-3742) YARN RM will shut down if ZKClient creation times out

[ https://issues.apache.org/jira/browse/YARN-3742?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15949251#comment-15949251 ] Karthik Kambatla commented on YARN-3742: I prefer this change over my recommendations on YARN-2814.

[jira] [Commented] (YARN-6407) Improve and fix locks of RM scheduler

[ https://issues.apache.org/jira/browse/YARN-6407?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15949242#comment-15949242 ] Karthik Kambatla commented on YARN-6407: How is locality related to continuous scheduling? You

[jira] [Commented] (YARN-6407) Improve and fix locks of RM scheduler

[ https://issues.apache.org/jira/browse/YARN-6407?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15947955#comment-15947955 ] Karthik Kambatla commented on YARN-6407: On large clusters, we have been recommending our customers

[jira] [Updated] (YARN-3774) ZKRMStateStore should use CuratorOp

[ https://issues.apache.org/jira/browse/YARN-3774?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-3774: --- Summary: ZKRMStateStore should use CuratorOp (was: ZKRMStateStore should use Curator 3.0 and

[jira] [Assigned] (YARN-3774) ZKRMStateStore should use Curator 3.0 and avail CuratorOp

[ https://issues.apache.org/jira/browse/YARN-3774?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla reassigned YARN-3774: -- Assignee: Robert Kanter (was: Karthik Kambatla) Thanks Robert for expressing interest

[jira] [Assigned] (YARN-2415) Expose MiniYARNCluster for use outside of YARN

[ https://issues.apache.org/jira/browse/YARN-2415?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla reassigned YARN-2415: -- Assignee: Haibo Chen (was: Karthik Kambatla) Thanks [~haibochen] for expressing

[jira] [Updated] (YARN-6246) Identifying starved apps does not need the scheduler writelock

[ https://issues.apache.org/jira/browse/YARN-6246?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Karthik Kambatla updated YARN-6246: --- Attachment: YARN-6246.003.patch Thanks for the review, Miklos. Good catch. New patch fixes

[jira] [Commented] (YARN-5685) Non-embedded HA failover is broken

[ https://issues.apache.org/jira/browse/YARN-5685?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15945790#comment-15945790 ] Karthik Kambatla commented on YARN-5685: +1 > Non-embedded HA failover is broken >

[jira] [Commented] (YARN-6050) AMs can't be scheduled on racks or nodes

[ https://issues.apache.org/jira/browse/YARN-6050?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15945772#comment-15945772 ] Karthik Kambatla commented on YARN-6050: +1 on the branch-2 patch as well. > AMs can't be

[jira] [Commented] (YARN-6359) TestRM#testApplicationKillAtAcceptedState fails rarely due to race condition

[ https://issues.apache.org/jira/browse/YARN-6359?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15944186#comment-15944186 ] Karthik Kambatla commented on YARN-6359: +1 > TestRM#testApplicationKillAtAcceptedState fails

[jira] [Comment Edited] (YARN-6359) TestRM#testApplicationKillAtAcceptedState fails rarely due to race condition

[ https://issues.apache.org/jira/browse/YARN-6359?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15941129#comment-15941129 ] Karthik Kambatla edited comment on YARN-6359 at 3/24/17 9:04 PM: - Nit: I

[jira] [Commented] (YARN-6359) TestRM#testApplicationKillAtAcceptedState fails rarely due to race condition

[ https://issues.apache.org/jira/browse/YARN-6359?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15941129#comment-15941129 ] Karthik Kambatla commented on YARN-6359: Nit: I think we use 100 ms in other places; maybe we

[jira] [Commented] (YARN-5685) Non-embedded HA failover is broken

[ https://issues.apache.org/jira/browse/YARN-5685?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15941122#comment-15941122 ] Karthik Kambatla commented on YARN-5685: Have no recollection from my previous review. Looked

[jira] [Commented] (YARN-6050) AMs can't be scheduled on racks or nodes

[ https://issues.apache.org/jira/browse/YARN-6050?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15939146#comment-15939146 ] Karthik Kambatla commented on YARN-6050: +1 pending Jenkins. [~leftnoteasy] - do you want to take

  1   2   3   4   5   6   7   8   9   10   >