[
https://issues.apache.org/jira/browse/OAK-4285?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Michael Dürig updated OAK-4285:
-------------------------------
Attachment: OAK_4285.patch
While working on OAK-4312 I tentatively came up with a patch for this issue:
[^OAK_4285.patch].
However after further considerations I'm note convinced that this is the right
thing to do. Maybe we should do nothing here. I.e. no clean up of the segment
id tables at all. After all, these tables track segment ids that are references
from somewhere within the JVM. Whether or not the underlying segment exists or
has been reclaimed is a different concern. Maybe we should just add an explicit
not to that respect to {{SegmentTracker#getReferencedSegmentIds}}.
[~alex.parvulescu], [~frm], WDYT?
> Align cleanup of segment id tables with the new cleanup strategy
> -----------------------------------------------------------------
>
> Key: OAK-4285
> URL: https://issues.apache.org/jira/browse/OAK-4285
> Project: Jackrabbit Oak
> Issue Type: Task
> Components: segment-next
> Reporter: Michael Dürig
> Assignee: Alex Parvulescu
> Labels: cleanup, gc
> Fix For: 1.6
>
> Attachments: OAK_4285.patch
>
>
> We need to align cleanup of the segment id tables with the new "brutal"
> strategy introduced with OAK-3348. That is, we need to remove those segment
> id's from the segment id tables whose segment have actually been gc'ed.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
