[
https://issues.apache.org/jira/browse/YARN-1618?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13880725#comment-13880725
]
Vinod Kumar Vavilapalli commented on YARN-1618:
-----------------------------------------------
Sorry Jian is away and I'm tied up with other things.
The point about saving app before scheduler acknowledges is a known issue. If
that is the only issue, we can close as a duplicate of YARN-1507 which already
exists.
I'll need to look carefully but this is why I think we did this. Because the
save happens asynchronously, an app may get killed while the app is persisted
but not acknowledged. That's the reason why it moves to FINAL_SAVING state
first. Obviously, that also means that there are cases like this when the
original save request hasn't reached the state-store yet.
> Applications transition from NEW to FINAL_SAVING, and try to update
> non-existing entries in the state-store
> -----------------------------------------------------------------------------------------------------------
>
> Key: YARN-1618
> URL: https://issues.apache.org/jira/browse/YARN-1618
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: resourcemanager
> Affects Versions: 2.2.0
> Reporter: Karthik Kambatla
> Assignee: Karthik Kambatla
> Priority: Blocker
> Attachments: yarn-1618-1.patch
>
>
> YARN-891 augments the RMStateStore to store information on completed
> applications. In the process, it adds transitions from NEW to FINAL_SAVING.
> This leads to the RM trying to update entries in the state-store that do not
> exist. On ZKRMStateStore, this leads to the RM crashing.
> Previous description:
> ZKRMStateStore fails to handle updates to znodes that don't exist. For
> instance, this can happen when an app transitions from NEW to FINAL_SAVING.
> In these cases, the store should create the missing znode and handle the
> update.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
