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 "Chetan Mehrotra (JIRA)" <ji...@apache.org> on 2017/12/11 05:46:00 UTC

[jira] [Commented] (OAK-7042) Pin async indexer on cluster leader

    [ https://issues.apache.org/jira/browse/OAK-7042?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16285527#comment-16285527 ] 

Chetan Mehrotra commented on OAK-7042:
--------------------------------------

+1. Looks well tested!

> Pin async indexer on cluster leader
> -----------------------------------
>
>                 Key: OAK-7042
>                 URL: https://issues.apache.org/jira/browse/OAK-7042
>             Project: Jackrabbit Oak
>          Issue Type: Improvement
>          Components: lucene
>            Reporter: Vikas Saurabh
>            Assignee: Vikas Saurabh
>            Priority: Blocker
>             Fix For: 1.7.13, 1.8
>
>
> Currently, async indexer creates some data locally to declare which index files have been deleted from the indexer. This information is used by active deletion's purge logic (to be invoked via jmx).
> Since, async indexer sets its task to scheduler to run as singleton on the cluster (not pinning to anywhere), then the active deletion jmx execution requires the knowledge where async indexer is scheduled.
> To avoid that uncertainty, it's better to instead pin the async indexer on cluster leader and document that active deletion jmx also needs to be invoked on cluster leader.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)