You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@jackrabbit.apache.org by "Julian Reschke (JIRA)" <ji...@apache.org> on 2016/12/01 07:36:58 UTC

[jira] [Commented] (JCR-3943) [jackrabbi-aws-ext] Data inconsistency due to race condition during async uploads

    [ https://issues.apache.org/jira/browse/JCR-3943?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15711197#comment-15711197 ] 

Julian Reschke commented on JCR-3943:
-------------------------------------

This bug has been open since January, but is marked "critical".

[~amitjain] - is it really critical? Can it be moved to a later release?

> [jackrabbi-aws-ext] Data inconsistency due to race condition during async uploads
> ---------------------------------------------------------------------------------
>
>                 Key: JCR-3943
>                 URL: https://issues.apache.org/jira/browse/JCR-3943
>             Project: Jackrabbit Content Repository
>          Issue Type: Bug
>          Components: jackrabbit-data
>            Reporter: Amit Jain
>            Assignee: Amit Jain
>            Priority: Critical
>             Fix For: 2.13.5, 2.14
>
>         Attachments: coredata14Nov5Dec.txt
>
>
> There is a race condition when {{LocalCache}} is used where if an upload ({{file_u}}) has entered the cache but not the {{AsyncUploadCache}} and a simultaneous PurgeJob is running, then the uploaded file {{file_u}} can be purged from the cache.
> When the async job ultimately runs it fails silently (S3 client fails to calculate the hash because of the missing file), thus leaving dangling references in the node store as well as the {{AsyncUploadCache}}.



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