[jira] [Commented] (SLING-7039) Clean up jobs in state dropped and errors

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

[jira] [Commented] (SLING-7039) Clean up jobs in state dropped and errors

JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/SLING-7039?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16121255#comment-16121255 ]

Karl Pauls commented on SLING-7039:
-----------------------------------

[~cziegeler], I see your point. Basically, I was interpreting the period mostly as a limit on the scanning while you see it as a user facing promise for the max age of stored events. Makes sense :-).

I'll change it to be:

{code:java}
        if (this.configuration.getHistoryCleanUpRemovedJobs() > 0 &&
                schedulerRuns % 60 == 1) {
            Calendar removeDate = Calendar.getInstance();
            removeDate.add(Calendar.MINUTE, - this.configuration.getHistoryCleanUpRemovedJobs());
            this.historyCleanUpRemovedJobs(removeDate);
        }
{code}

which makes it so that we scan every hour and delete everything older than the configured period (which should be good enough).


> Clean up jobs in state dropped and errors
> -----------------------------------------
>
>                 Key: SLING-7039
>                 URL: https://issues.apache.org/jira/browse/SLING-7039
>             Project: Sling
>          Issue Type: Improvement
>          Components: Extensions
>            Reporter: Carsten Ziegeler
>            Assignee: Karl Pauls
>             Fix For: Event 4.3.0
>
>         Attachments: SLING-7039.patch
>
>
> Currently, whenever a job is removed by a user (or client code), the job is kept as a dropped job - regardless of whether a history should be kept for these types of jobs. The same happens if the job can't be scheduled (error).
> This information is useful for trouble shooting (to find out if a particular job has been created at all or not).
> Now, as the job impl is keeping track of these things, I think it also should be the job impl which cleans this up. The configuration for this would be a time period during which these jobs are kept (like 48h or something). All older jobs are removed.
> Or in other words: the job impl should periodically run the HistoryCleanUpTask to remove jobs in state DROPPED and ERROR which are older than the configured time.
> I think the time would be a global configuration of the job handling, defaulting to 48h



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