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 "Marcel Reutegger (Jira)" <ji...@apache.org> on 2022/11/14 16:06:00 UTC
[jira] [Updated] (OAK-9992) Introduce limit for oak.removeCollisions()
[ https://issues.apache.org/jira/browse/OAK-9992?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Marcel Reutegger updated OAK-9992:
----------------------------------
Summary: Introduce limit for oak.removeCollisions() (was: ntroduce limit for oak.removeCollisions())
> Introduce limit for oak.removeCollisions()
> ------------------------------------------
>
> Key: OAK-9992
> URL: https://issues.apache.org/jira/browse/OAK-9992
> Project: Jackrabbit Oak
> Issue Type: Improvement
> Components: run
> Reporter: Marcel Reutegger
> Assignee: Marcel Reutegger
> Priority: Minor
>
> Currently the utility method in oak.removeCollisions() attempts to remove all collision markers for a given clusterId. With many collisions this can be a rather expensive operation that is difficult to schedule for MongoDB. I've seen cases where removing all collisions during DocumentNodeStore initialization was blocked for a very long time on the update call to MongoDB. On the MongoDB side the operation was retried many times because it conflicted with other concurrent updates on the same document.
> The method should have a third parameter that limits the number of collisions to remove in one call.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)