[
https://issues.apache.org/jira/browse/YARN-4834?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jason Lowe updated YARN-4834:
-
Target Version/s: 2.7.4
+1 to using the session ID to track the processes for a container. Ideally if
we're using cgroups we should use that instead, but in the interim this would
be a significant improvement.
I'm not thrilled with keeping the double-pass logic leftover from the
parent-child tree calculations. The session ID approach would only require a
single pass. However that's a significant rewrite of the code, and I can
appreciate keeping the changes to a minimum to fix this bug. We can file a
followup JIRA to simplify the logic.
+1 lgtm. Will commit this tomorrow if there are no objections.
> ProcfsBasedProcessTree doesn't track daemonized processes
> -
>
> Key: YARN-4834
> URL: https://issues.apache.org/jira/browse/YARN-4834
> Project: Hadoop YARN
> Issue Type: Bug
> Components: nodemanager
>Affects Versions: 3.0.0, 2.7.2
>Reporter: Nathan Roberts
>Assignee: Nathan Roberts
> Attachments: YARN-4834.001.patch
>
>
> Currently the algorithm uses ppid from /proc//stat which can be 1 if a
> child process has daemonized itself. This causes potentially large processes
> from not being monitored.
> session id might be a better choice since that's what we use to signal the
> container during teardown.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
-
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
