jackrabbit-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jukka Zitting (JIRA)" <j...@apache.org>
Subject [jira] Commented: (JCR-1722) Data Store backup: need a way to delay deleting files (garbage collection)
Date Wed, 08 Oct 2008 12:05:44 GMT

    [ https://issues.apache.org/jira/browse/JCR-1722?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12637886#action_12637886
] 

Jukka Zitting commented on JCR-1722:
------------------------------------

AFAIK the garbage collection is only run when explicitly invoked. As long as neither garbage
collection nor such backup mechanisms are automatically run by the repository I think it would
be better to handle such coordination on a level above the repository.

> Data Store backup: need a way to delay deleting files (garbage collection)
> --------------------------------------------------------------------------
>
>                 Key: JCR-1722
>                 URL: https://issues.apache.org/jira/browse/JCR-1722
>             Project: Jackrabbit
>          Issue Type: New Feature
>          Components: jackrabbit-core
>            Reporter: Thomas Mueller
>            Assignee: Thomas Mueller
>             Fix For: 1.5
>
>
> During a backup, objects in the data store must not be modified or deleted. Otherwise
the backup is consistent.
> Objects in the data store are never modified. The only problem is if objects are deleted
while the backup is running. The data store garbage collection is the only place where objects
are deleted.
> There should be a way for a backup tool to 'flag' the data store (file data store and
database data store) to not delete entries. For the file data store, this could be done by
placing a specially named file in the root folder. For the database data store a specially
named entry could be used. Another idea is to use an in-memory state only, however this solution
doesn't work if multiple repositories share the same data store.
> To support shared data stores, the specially named file name / object name should consist
of a fixed part (for example the prefix 'stopdelete') and a unique id (for example a UUID).

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message