[midPoint] midPoint crash when validity scanner has too much objects

Pavol Mederly mederly at evolveum.com
Mon Jul 5 11:03:46 CEST 2021


Hello Pascal,

tasks in midPoint should not load all processed objects into memory at 
once. However, there might be some memory leaks hidden deep in code - 
from time to time we find (and fix) one of these.

You could browse our jira for recently fixed memleaks and see if an 
upgrade from your version would help (btw, what version of mP do you use?).

If nothing would apply, you could take a memory dump, and analyze it 
using e.g. Eclipse Memory Analyzer <https://www.eclipse.org/mat/> and 
open a Jira issue providing the data.

Best regards,

-- 
Pavol Mederly
Software developer
evolveum.com

On 05/07/2021 10:54, Pascal PÉRICHON via midPoint wrote:
> Hi,
>
> We have 25000 student accounts to close (lifecycleState = archived) 
> because date validTo is over.
>
> Validity scanner have to work on 380000 objects and midpoint crash 
> violently with an out of memory (heap size).
>
> We increase the heap size until 42 Go but it crashed at 100000 objects.
>
> midPoint seems to load all the objects in memory ?
>
> Is there any workaround to avoid midpoint to crash (tasks to recompute 
> and reconciliate accounts ?)
>
> thanks
>
> _______________________________________________
> midPoint mailing list
> midPoint at lists.evolveum.com
> https://lists.evolveum.com/mailman/listinfo/midpoint
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.evolveum.com/pipermail/midpoint/attachments/20210705/98c59f45/attachment.htm>


More information about the midPoint mailing list