Just Upgraded to 11.3 and now 'OpenJDK Platform Binary' is using almost all my memory

[CW-GIS]

Is OpenJDK Platform Binary anything to do with Enterprise. I don't recall seeing this at all before upgrading to 11.3 the other day. It is just eating up a ton of resources.

[CW-GIS]

We just upped our Azure machine and that stopped it from crashing, but the memory is still high

[CW-GIS]

Well we doubled the memory in our machine about a year ago and it has been going fine. But just recently our data store has been having issues and it might be this coming back to haunt us again.

[DanCamboia]

Just to give an update. Uninstalling the Cumulative Update worked for about a week and then the problem started happening again. After installing the November Cumulative Update things were a little better, but not completely resolved. Since the December Cumulative Update I'm no longer experiencing this issue. 

[MoMahmoud]

Has anyone seen the same behaviour on the datastore?

[NatC42]

Hi @MoMahmoud.... We've been having issues with our hosting server logging database connection issues with our ArcGIS Data Store on another machine.  That machine is a Windows Server 2019 with 32GB Memory and 25GB Memory is constantly eaten up by the java.exe for the OpenJDK Platform binary.

Do we really need to double the memory on the server to account for this java application or can we mitigate the issue another way?

Attached is a screen shot of the Server Resources... 

[NatC42]

Looks like this issue was documented as a bug for the GeoAnalytics Server for previous versions (10.7 - 10.9.1) and it was recommended to just kill that OpenJDK process.  https://support.esri.com/en-us/bug/suspected-memory-leak-in-arcgis-geoanalytics-server-bug-000139823 

I would really like to know what piece of ArcGIS Enterprise needs that OpenJDK for GeoAnalytics it was recommended it wasn't needed. Is it any ArcGIS Server installed instance hosting services, the ArcGIS Data Store Service, Portal, Hosting Server, etc.  

[DurmusCesar]

The same here, I do see tones of memory leak message in Tomcat logs, ESRI reps better jump to this post explain what caused this in 11.3 version and offer some resolution on this to me

One thing I could see tones of memory leak messages in Tomcat logs managed by Portal 11.3

....own command was received via the shutdown port. Stopping the Server instance..... To prevent a memory leak, the JDBC Driver has been forcibly unregistered.

.....but has failed to stop it. This is very likely to create a memory leak. Stack trace of thread:=....

Not sure how current ArcGIS Portal/Server/RDS (enterprise) 11.3 portal managing Tomcat? Do ESRI have any suggestion other than bump up the memory (indefinitely???), restart every single second, min, hour (so that you get some relief for sometime till it picks up again) any suggestions and/or more of fix on ESRI side (11.2 OK 11.3 not OK in this regard, and similar observations with 10.0 and 10.1 in the past? ON/OFF issue???