You are viewing a plain text version of this content. The canonical link for it is here.
Posted to oak-issues@jackrabbit.apache.org by "Michael Dürig (JIRA)" <ji...@apache.org> on 2017/08/08 15:36:00 UTC
[jira] [Commented] (OAK-6520) Improve tail compactions resilience
when base state cannot be determined
[ https://issues.apache.org/jira/browse/OAK-6520?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16118480#comment-16118480 ]
Michael Dürig commented on OAK-6520:
------------------------------------
Looking at the code it seems like the only case missing is the one where the record id determined for the base state points to a segment that does not exist.
> Improve tail compactions resilience when base state cannot be determined
> ------------------------------------------------------------------------
>
> Key: OAK-6520
> URL: https://issues.apache.org/jira/browse/OAK-6520
> Project: Jackrabbit Oak
> Issue Type: Improvement
> Components: segment-tar
> Reporter: Michael Dürig
> Assignee: Michael Dürig
> Labels: compaction, gc
> Fix For: 1.8, 1.7.7
>
>
> This is a follow up to OAK-3349: in tail compaction the {{FileStore.GarbageCollector.getBase()}} might fail to determine the base state to rebase onto. In this case we should fall back to full compaction and report (log, JMX) the problem and its exact cause.
> Failing to determine the base state might be caused by a missing or invalid {{gc.log}} file or a invalid or missing record id for the base state being recorded in {{gc.log}}. None of these cases should impact system stability.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)