You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@kudu.apache.org by "Todd Lipcon (JIRA)" <ji...@apache.org> on 2016/03/09 23:31:41 UTC

[jira] [Resolved] (KUDU-218) Should exercise case where a duplicated insert was flushed from neither store

     [ https://issues.apache.org/jira/browse/KUDU-218?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Todd Lipcon resolved KUDU-218.
------------------------------
    Resolution: Cannot Reproduce
      Assignee: Todd Lipcon

When i implemented KUDU-969 I investigated this case and the TODO related to this is no longer relevant.

> Should exercise case where a duplicated insert was flushed from neither store
> -----------------------------------------------------------------------------
>
>                 Key: KUDU-218
>                 URL: https://issues.apache.org/jira/browse/KUDU-218
>             Project: Kudu
>          Issue Type: Bug
>          Components: tablet, tserver
>    Affects Versions: M4
>            Reporter: Todd Lipcon
>            Assignee: Todd Lipcon
>
> David and I were conversing on IRC and thinking about the case where a duplicated update ends up in both the "old" and "new" rowsets of a compaction, but the server crashes before the compaction finishes. In that case, it would show up as "unflushed" in both the before and after targets during bootstrap. I added a FATAL for this case in the bootstrap code and it never triggered, implying that we probably lack test coverage for this important case.



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