[
https://issues.apache.org/jira/browse/YARN-6153?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
kyungwan nam updated YARN-6153:
-------------------------------
Attachment: YARN-6153-branch-2.8.patch
I'm uploading the patch for branch-2.8.
1. in the testRMAppAttemptFailuresValidityInterval, using the systemClock has
been replaced with Thread.sleep.
by following, the time to check the validity interval is no longer the
systemClock in RMAppImpl.
{code}
- private int getNumFailedAppAttempts() {
+ public int getNumFailedAppAttempts() {
int completedAttempts = 0;
- long endTime = this.systemClock.getTime();
// Do not count AM preemption, hardware failures or NM resync
// as attempt failure.
for (RMAppAttempt attempt : attempts.values()) {
if (attempt.shouldCountTowardsMaxAttemptRetry()) {
- if (this.attemptFailuresValidityInterval <= 0
- || (attempt.getFinishTime() > endTime
- - this.attemptFailuresValidityInterval)) {
- completedAttempts++;
- }
+ completedAttempts++;
}
}
{code}
2. in the testAMRestartNotLostContainerAfterAttemptFailuresValidityInterval,
the timeout value has been increased to 40 seconds.
currently, YARN-4807 is not yet included in the branch-2.8. I think that’s why
the timeout happens.
> keepContainer does not work when AM retry window is set
> -------------------------------------------------------
>
> Key: YARN-6153
> URL: https://issues.apache.org/jira/browse/YARN-6153
> Project: Hadoop YARN
> Issue Type: Bug
> Components: resourcemanager
> Affects Versions: 2.7.1
> Reporter: kyungwan nam
> Assignee: kyungwan nam
> Fix For: 2.8.0, 3.0.0-alpha3
>
> Attachments: YARN-6153.001.patch, YARN-6153.002.patch,
> YARN-6153.003.patch, YARN-6153.004.patch, YARN-6153.005.patch,
> YARN-6153.006.patch, YARN-6153-branch-2.8.patch
>
>
> yarn.resourcemanager.am.max-attempts has been configured to 2 in my cluster.
> I submitted a YARN application (slider app) that keepContainers=true,
> attemptFailuresValidityInterval=300000.
> it did work properly when AM was failed firstly.
> all containers launched by previous AM were resynced with new AM (attempt2)
> without killing containers.
> after 10 minutes, I thought AM failure count was reset by
> attemptFailuresValidityInterval (5 minutes).
> but, all containers were killed when AM was failed secondly. (new AM attempt3
> was launched properly)
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
