[jira] [Commented] (MESOS-2473) Failure to recover because of freezer timeout should not suggest removing meta data

["Marco Massenzio (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/MESOS-2473?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14596906#comment-14596906 ] 

Marco Massenzio commented on MESOS-2473:
----------------------------------------

Unfortunately, it appears we won't get to do this in time for 0.23 - nothing happened for a couple of months and it's not critical for release.

> Failure to recover because of freezer timeout should not suggest removing meta data
> -----------------------------------------------------------------------------------
>
>                 Key: MESOS-2473
>                 URL: https://issues.apache.org/jira/browse/MESOS-2473
>             Project: Mesos
>          Issue Type: Improvement
>          Components: isolation
>    Affects Versions: 0.22.0
>            Reporter: Ian Downes
>              Labels: twitter
>
> A more appropriate action should be suggested, e.g., manually kill the processes in cgroup <xxx> because the slave will still attempt to clean up orphans and hit the same code path.
> {noformat}
> I0310 23:04:23.961019 32342 slave.cpp:3321] Current usage 35.87%. Max allowed age: 3.789365411204225days
> Failed to perform recovery: Collect failed: Timed out after 1mins
> To remedy this do as follows:
> Step 1: rm -f /var/lib/mesos/meta/slaves/latest
>         This ensures slave doesn't recover old live executors.
> Step 2: Restart the slave.
> Slave Exit Status: 1
> {noformat}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)