[jira] [Commented] (OAK-1647) AsyncIndexUpdateTask creating too many checkpoint

["Chetan Mehrotra (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/OAK-1647?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13954972#comment-13954972 ] 

Chetan Mehrotra commented on OAK-1647:
--------------------------------------

[~alexparvulescu] Can you have a look

With current frequency of 5 sec 1 checkpoint entry would be created for every run i.e. 720 entries per clutser node/per hour. And they can only be removed after 1 hour (default expiry time)

If such a change is not possible then this would cause issue depending on how checkpoints are stored. In case of DocumentNodeStore the checkpoint entry is currently saved in a single Mongo document as a sub property. A Mongo document can store 16 MB of data so should be able to store a bit.

This was done with the assumption that checkpoints are not created in large numbers :(. That implementation can be changed to say store checkpoint data as independent collection. 


> AsyncIndexUpdateTask creating too many checkpoint
> -------------------------------------------------
>
>                 Key: OAK-1647
>                 URL: https://issues.apache.org/jira/browse/OAK-1647
>             Project: Jackrabbit Oak
>          Issue Type: Improvement
>          Components: core
>            Reporter: Chetan Mehrotra
>            Assignee: Chetan Mehrotra
>            Priority: Minor
>             Fix For: 0.20
>
>
> AsyncIndexUpdateTask currently creates a checkpoint before [1] proceeding to index. If the indexer is already running then also the checkpoint gets created as that check is performed later. 
> Would it be possible to create checkpoint only after checking if the async index update is already running or not? 
> [1] https://github.com/apache/jackrabbit-oak/blob/trunk/oak-core/src/main/java/org/apache/jackrabbit/oak/plugins/index/AsyncIndexUpdate.java#L117



--
This message was sent by Atlassian JIRA
(v6.2#6252)