You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@sling.apache.org by "Stefan Egli (JIRA)" <ji...@apache.org> on 2016/05/02 11:01:12 UTC
[jira] [Resolved] (SLING-5705) ChunkCleanUpTask to not run
concurrently by default (configurable)
[ https://issues.apache.org/jira/browse/SLING-5705?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Stefan Egli resolved SLING-5705.
--------------------------------
Resolution: Fixed
added {{scheduler.concurrent==false}} in http://svn.apache.org/viewvc?rev=1741940&view=rev
> ChunkCleanUpTask to not run concurrently by default (configurable)
> ------------------------------------------------------------------
>
> Key: SLING-5705
> URL: https://issues.apache.org/jira/browse/SLING-5705
> Project: Sling
> Issue Type: Improvement
> Components: Servlets
> Affects Versions: Servlets Post 2.3.10
> Reporter: Stefan Egli
> Assignee: Stefan Egli
> Fix For: Servlets Post 2.3.12
>
>
> The {{ChunkCleanUpTask}} is scheduled to run every 12 hours. Typically it would end rather very quickly. But in SLING-5701 and SLING-5702 situations have been seen where it was entering an endless loop, thus would still run after 12 hours.
> Arguably, that is not a realistic scenario. But what happened then is that the scheduler started another {{ChunkCleanUpTask}} after 12 hours and so forth, until all the scheduler's thread pool threads were used. Thus this blocked the scheduler from operating after like 2 days.
> While certainly SLING-5701 and SLING-5702 are the root cause of this, a simple additional change should be to mark the {{ChunkCleanUpTask}} as {{scheduler.concurrent==false}} - which means it would not start a second task after 12 hours. Again, this shouldn't normally be of any issue, but for paranoia reasons and logically this task should never really have to be started concurrently.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)