Submitted jobs go to error status

We have an issue where about every 7 to 10 days our submitted jobs go into an error status. We have to recycle the services on our UBE server to allow jobs to process. The Metadata kernel log shows a memory error and I have an open SR with Oracle. Oracle suggested we change the JDE.INI file seciton [JDE JVM] for DefaultOptions= -Xms128m -Xmx256m
This setting is suppose to increase the memory available to the Metadata kernel. After we put this setting in no UBEs could be submitted and the JDE Scheduler would not start. I am still waiting on Oracle to give me another setting or explain to me why this is happening. Has anyone ever experienced this issue or know a fix for it? Thanks in advance.

We are using the following with no issues.



DefaultOptions=3D-XX:-SafepointPolling

MetadataVMOptions=3D-Xms256m -Xmx512m

InProcessJVMHome=3D

RemoteDebugPort=3D-1

VMTrace=3D0

XMLPVMOPTIONS=3D-Djava.compiler=3DNONE -Xmx256m



Dave Rammer

Sheboygan County IS



E900 TR 8.98.2.1.CPU OAS 10.1.3.4.0 HP-UX 11.31 Oracle DB 11.1.0.7.0

Hi,

We too have the same problems before and ORACLE support suggested this jde.ini setting on our enterpriseone server:

[JDE JVM]
DefaultOptions=-server -Xrs -Xms128m -Xmx256m

"-server -Xrs" - check docid 663276.1
"-Xms128m -Xmx256m" - check docid 790848.1

Anyways we also have our jde service restart daily during offpeak hours (batch scheduler) to clean the memory it uses...so far it worked for us.

I guess these kernels are still not efficient as it eats way too much (leaking) memory to my liking... hope they really find ways to improve on it.

Hope this helps.