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 "Julian Reschke (JIRA)" <ji...@apache.org> on 2017/01/17 12:39:26 UTC

[jira] [Reopened] (OAK-5461) [BlobGC] BlobIdTracker remove() should merge generations

     [ https://issues.apache.org/jira/browse/OAK-5461?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Julian Reschke reopened OAK-5461:
---------------------------------

The latest change appears to introduce a reproducible hangup on Windows for {{BlobIdTrackerStoreTest}}

> [BlobGC] BlobIdTracker remove() should merge generations
> --------------------------------------------------------
>
>                 Key: OAK-5461
>                 URL: https://issues.apache.org/jira/browse/OAK-5461
>             Project: Jackrabbit Oak
>          Issue Type: Bug
>          Components: blob
>            Reporter: Amit Jain
>            Assignee: Amit Jain
>             Fix For: 1.5.18, 1.6
>
>
> On a new setup if the blob gc is executed before taking a snapshot then the blobs ids are retrieved from the datastore and GC removes are synchronized with this retrieved file. But these removed blob ids may also be present in the tracker generation and therefore may get resurrected on next snapshot. This will happen as no snapshot has taken place which is possible if the {{blobGcMaxAgeInSecs}} is configured to be less than the {{blobTrackSnapshotIntervalInSecs}}.
> The effect of the above situation would be that the removed blob ids would always turn up as GC candidates until the tracker is reset and log a warning during GC as they are not present anymore. 



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