[jira] [Updated] (OAK-3475) Confusing SNFE whith oak-run debug

["Michael Dürig (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/OAK-3475?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Michael Dürig updated OAK-3475:
-------------------------------
    Fix Version/s: 1.3.8

> Confusing SNFE whith oak-run debug
> ----------------------------------
>
>                 Key: OAK-3475
>                 URL: https://issues.apache.org/jira/browse/OAK-3475
>             Project: Jackrabbit Oak
>          Issue Type: Improvement
>          Components: run
>            Reporter: Michael Dürig
>            Assignee: Michael Dürig
>              Labels: cleanup, gc, tooling
>             Fix For: 1.3.8
>
>
> Running 
> {{oak-run debug  /path/to/segmentStore}}
> can result in {{SNFE}} s being logged if the file store has been compacted before. This can happen even though the repository is actually consistent according to {{oak-run check}}. 
> The reason is {{debug}} traversing all node states of all record ids in all tar files. When a previous cleanup of a tar file did not reach 25% gain it will not clean up that file leaving behind segments possibly containing node states pointing to limbo. Those nodes would result in said {{SNFE}} of {{oak-run}} debug. But as those node states are not reachable from the head node state the repository is still consistent itself. 



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