Anyone else having issues with Jboss not using enough allocated memory to run properly or have been told that by GE support?
Currently have an issue where it allocates a lot but will not use more than a third and supposedly it is causing issues where we are having problems adding diagnosis and seeing Jboss 500 error messages.
UPDATE: GE support that they are seeing this issue with other sites since SP10 came out. So looks like we have another "enhancement" that was documented
We haven't run into this yet. We installed SP10 this past weekend. How much memory do you have allocated?
We started seeing a few random "out of memory" messages pop-up to users on their Citrix client after having already been live for 6-7 weeks on SP10 EMR9.8.
I hadn't connected it to any possible JBOSS issues, so thanks for the info. The JBOSS server itself has seemed normal and we'd been thinking it was related to a given Citrix server. Fortunately, the error doesn't happen too often.
Currently have 10Gb allocated of memory that the server configurator suggested.
Worked with GE tier 3 till around 10pm last night. Looks like it is an issue in the wrapper config with the +UseLargePages attribute. We disabled it and started working correctly and got up to the 2gb of ram used like it should. We then reenabled the LargePages and it worked also. Current theory is when it crashes, the wrapper config will restart jboss after 60 secs and that is not long enough for the memory to clear out the old instance of jboss and then it gets hung up around the 260Mb mark.
If we waited for the jboss to completely release the memory and then restarted, it would work properly.
I have a follow up call with GE at 1pm CST today to go over it again. Will update this after that call.
Also forgot to mention, preceding the issue with jboss, our problem searches would start to crash or lock up for long periods of time.
They think it could be something that is corrupting the cached indexes that jboss does.
Wondering if there is an update on this issue.
Due to the release notes published last night with the issue about September knowledge base we have postponed our upgrade but interested if there is a solution to this.
We are watching it right now. Nothing new as of now to report. Everything has been stable for the last day and seems to be running fine.