[
https://issues.apache.org/jira/browse/SLIDER-426?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14165367#comment-14165367
]
thomas liu commented on SLIDER-426:
-----------------------------------
In sliderClient.actionFreeze():
after checking application state is < YarnApplicationState.FINISHED, if the
application state is not in ‘YarnApplicationState.RUNNING’, then mark boolean
local variable forcekill to be true
Race condition here may happen:
1) The application was in non running state when querying, and however, moved
to running state when killing it.
Given we are killing the application by force anyway, there wouldn’t be bad
results
2) The application was in running state when querying, and however, moved to
non running state when killing it.
Current decision is to leave with this race condition
actionDestroy() is left as it is
> If an app is in ACCEPTED state then it cannot be stopped or destroyed
> ---------------------------------------------------------------------
>
> Key: SLIDER-426
> URL: https://issues.apache.org/jira/browse/SLIDER-426
> Project: Slider
> Issue Type: Bug
> Components: client
> Affects Versions: Slider 0.50
> Reporter: Sumit Mohanty
> Assignee: thomas liu
> Fix For: Slider 0.60
>
>
> {noformat}
> [yarn@c6401 bin]$ ./slider stop cl1
> ...
> 2014-09-13 15:14:24,357 [main] INFO client.RMProxy - Connecting to
> ResourceManager at c6401.ambari.apache.org/192.168.64.101:8050
> 2014-09-13 15:14:26,042 [main] WARN client.SliderClient - Exception while
> trying to terminate cl1:
> org.apache.slider.core.exceptions.BadClusterStateException: Application not
> running: application_1410619964429_0001 state=ACCEPTED
> 2014-09-13 15:14:26,051 [main] INFO util.ExitUtil - Exiting with status -1
> [yarn@c6401 bin]$ ./slider destroy cl1
> ...
> 2014-09-13 02:13:39,837 [main] INFO client.RMProxy - Connecting to
> ResourceManager at c6401.ambari.apache.org/192.168.64.101:8050
> 2014-09-13 02:13:41,687 [main] ERROR main.ServiceLauncher - cl1: Application
> Instance already running :applicationId { id: 1 cluster_timestamp:
> 1410570014723 } user: "yarn" queue: "default" name: "cl1" host: "N/A"
> rpc_port: -1 yarn_application_state: ACCEPTED trackingUrl:
> "http://c6401.ambari.apache.org:8088/proxy/application_1410570014723_0001/";
> diagnostics: "" startTime: 1410574245847 finishTime: 0
> final_application_status: APP_UNDEFINED app_resource_Usage {
> num_used_containers: 0 num_reserved_containers: 0 used_resources { memory: 0
> virtual_cores: 0 } reserved_resources { memory: 0 virtual_cores: 0 }
> needed_resources { memory: 0 virtual_cores: 0 } } originalTrackingUrl: "N/A"
> currentApplicationAttemptId { application_id { id: 1 cluster_timestamp:
> 1410570014723 } attemptId: 1 } progress: 0.0 applicationType:
> "org-apache-slider" applicationTags: "name: hbase" applicationTags: "version:
> 0.98.5-hadoop2" applicationTags: "description: apache hbase is the hadoop
> database, a distributed, scalable, big data store...."
> 2014-09-13 02:13:41,693 [main] INFO util.ExitUtil - Exiting with status 74
> {noformat}
> [~thomas_liu] can you discuss with [[email protected]] regarding this?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
