Using the LTS version is sadly not an option - I would definitely prefer 
that.
But with our amount of projects and builds it takes a couple of hours to 
even start Jenkins because the lazy-loading fix was not yet released at 
that time.

To give you a small example how ridiculous the "triggered-by-upstream" data 
can be please take a look to the following job:
http://jenkins.willowgarage.com:8080/job/ros-groovy-pr2-object-manipulation_binarydeb_precise_amd64/98/consoleFull

Since I don't have any experience with the Jenkins sources myself I am 
pretty sure that my attempt to fix the issue would take a much longer time 
that any Jenkins developer and is likely not nearly as clean as it should 
be to be accepted upstream.

- Dirk


On Thursday, January 17, 2013 6:49:08 PM UTC-8, Mark Waite wrote:
>
> If it is urgent to disable the feature and the feature was added in 1.482, 
> you could consider installing the long term release, which is curently 
> 1.480.2.
>
> I'm impressed that you're running 5000 jobs with 30 executors and not 
> using the long term support release.  The tip of Jenkins development is 
> surprisingly stable, but with 5000 jobs, even a little instability seems 
> like it could be a major disruption.
>
> You could also fork the Jenkins repository, prepare the change you're 
> suggesting, and submit a pull request.  I did something like that with the 
> git plugin and found the process surprisingly friendly, even for me.
>
> Thanks,
> Mark Waite
>
>   ------------------------------
> *From:* "[email protected] <javascript:>" 
> <[email protected]<javascript:>
> >
> *To:* [email protected] <javascript:> 
> *Sent:* Thursday, January 17, 2013 5:09 PM
> *Subject:* Logging UpstreamCause floods Jenkins - need to optionally 
> disable that feature
>  
> I raised this question already on the users mailing list and filled a bug 
> reports months ago.
> Since I got no response yet and consider my issue a pretty severe one I am 
> reposting it to the dev list.
>
> In version 1.482 the feature "Report root causes of UpstreamCause in log 
> and status pages" has been added. In certain scenarios (as stated 
> below) this is absolutely not feasible because the amount of data logged 
> per build might become dozens of megabytes (sometimes hundreds). The result 
> is that the jobs folder grows for several thousand builds in tens of 
> gigabytes (within a couple of hours) which lets Jenkins hit memory limits 
> and become unusable (even on a machine with 64GB of memory).
>
> Some more words on the scenario which shows that problem. We have a 
> Jenkins instance with 30 executors, and about 5000 jobs. I think the 
> specific thing is that these jobs are not independent (or slightly 
> connected) but have a lot of up/downstream relationships. The problem is 
> that when Jenkins hits one of the leaf jobs the list of hierarchic causes 
> which triggered that job is tens of megabytes long (I am not attaching a 
> full log I guess the content is pretty obvious). On the one hand because 
> the nesting level is very high and on the other hand since there are 
> several paths through the dependency graph.
>
> [Conceptionally it looks even weird that Jenkins tries to keep all these 
> data in memory - but that might be more difficult to fix in the short-term.]
>
> So there is an urgent need to optionally disable that feature. As 
> mentioned I already filled a ticket about that: 
> https://issues.jenkins-ci.org/browse/JENKINS-15747
> Since this is basically making Jenkins unusable in such scenarios can this 
> be handled with priority?
>
> - Dirk
>
>
>   

Reply via email to