You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@sling.apache.org by "Chetan Mehrotra (JIRA)" <ji...@apache.org> on 2017/05/01 07:29:04 UTC
[jira] [Commented] (SLING-6797) Deprecate scheduling a job on a
specific Sling instance
[ https://issues.apache.org/jira/browse/SLING-6797?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15990645#comment-15990645 ]
Chetan Mehrotra commented on SLING-6797:
----------------------------------------
[~cziegeler] I think a better variant of this feature would be to allow pinning a job to instance having some tag defined. This could be based on [InstanceDesciption#properties|https://sling.apache.org/apidocs/sling7/index.html?org/apache/sling/discovery/InstanceDescription.html].
So a usecase here is that in a cluster setup we have certain instances configured for front end request while some instances are configured for background processing. In such a setup I would like the indexing job to be processed on one of the instances which are meant for background processing.
I beling Sling Discovery would allow me to "tag" such Sling instances and surface the tag value as part of InstanceDescription and then Sling Scheduler can use this to select the instance which would process such a job
[~cziegeler] [~egli] Thoughts? Possibly we should deal this in a new issue as this feature was specifically for jobs pinned to specific sling id
/cc [~ianeboston]
> Deprecate scheduling a job on a specific Sling instance
> -------------------------------------------------------
>
> Key: SLING-6797
> URL: https://issues.apache.org/jira/browse/SLING-6797
> Project: Sling
> Issue Type: Improvement
> Components: Commons
> Reporter: Carsten Ziegeler
> Assignee: Carsten Ziegeler
> Fix For: Commons Scheduler 2.5.4
>
>
> Currently it's possible to schedule a job on a specific instance by specifying the exact instance id. This feature is questionable, especially with instances coming and going it's hard to know the id in advance.
> It's better to rely on a good mechanism to distribute the scheduled jobs amongst availabe instances or use the leader
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)