[
https://issues.apache.org/jira/browse/YARN-1408?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13901288#comment-13901288
]
Sunil G commented on YARN-1408:
-------------------------------
An update on this patch:
For preemption policy, getLiveConatiners() is used to get the running
containers.
But container is added to this list when it is in ALLOCATED state itself.
So in some race conditions, it is possible that a container can get KILLED by
preemption even before it reach RUNNING state.
This scenario can be avoided if we can skip such containers which didnt reach
the RUNNING state during preemption.
May be in the following cycles this container will reach RUNNING state and the
can be considered for preemption.
[~devaraj.k] please check this updated patch and let know your opinion.
> Preemption caused Invalid State Event: ACQUIRED at KILLED and caused a task
> timeout for 30mins
> ----------------------------------------------------------------------------------------------
>
> Key: YARN-1408
> URL: https://issues.apache.org/jira/browse/YARN-1408
> Project: Hadoop YARN
> Issue Type: Bug
> Components: resourcemanager
> Affects Versions: 2.2.0
> Reporter: Sunil G
> Fix For: 2.2.0
>
> Attachments: Yarn-1408.1.patch, Yarn-1408.2.patch, Yarn-1408.patch
>
>
> Capacity preemption is enabled as follows.
> * yarn.resourcemanager.scheduler.monitor.enable= true ,
> *
> yarn.resourcemanager.scheduler.monitor.policies=org.apache.hadoop.yarn.server.resourcemanager.monitor.capacity.ProportionalCapacityPreemptionPolicy
> Queue = a,b
> Capacity of Queue A = 80%
> Capacity of Queue B = 20%
> Step 1: Assign a big jobA on queue a which uses full cluster capacity
> Step 2: Submitted a jobB to queue b which would use less than 20% of cluster
> capacity
> JobA task which uses queue b capcity is been preempted and killed.
> This caused below problem:
> 1. New Container has got allocated for jobA in Queue A as per node update
> from an NM.
> 2. This container has been preempted immediately as per preemption.
> Here ACQUIRED at KILLED Invalid State exception came when the next AM
> heartbeat reached RM.
> ERROR
> org.apache.hadoop.yarn.server.resourcemanager.rmcontainer.RMContainerImpl:
> Can't handle this event at current state
> org.apache.hadoop.yarn.state.InvalidStateTransitonException: Invalid event:
> ACQUIRED at KILLED
> This also caused the Task to go for a timeout for 30minutes as this Container
> was already killed by preemption.
> attempt_1380289782418_0003_m_000000_0 Timed out after 1800 secs
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
