Starting services always takes two attempts?

MrBlack · Post by **MrBlack** » Fri Jun 10, 2016 3:46 pm

Whenever I go to start the services, it will hang forever (45+ mins) and then I lose paitiants and I kill it, and search for the process and kill it. Then I issue the command a second time and it starts up within a couple of minutes. Can anyone offer guidance where to start looking as to why this is???

Code: Select all

/opt/IBM/InformationServer/ASBServer/bin/MetadataServer.sh run

About my environment:
Linux
Two machines, one for services, one for engine
Version 11.3
Is there anything else that would be helpful to know?

JRodriguez · Post by **JRodriguez** » Fri Jun 10, 2016 4:46 pm

I would check if last time was stopped clean by looking at the logs and searching for any Java process running in the server. If any present that will be the culprit of your issue..

eostic · Post by **eostic** » Sun Jun 12, 2016 2:57 pm

And also check out the work load manager....maybe there isnt enough cpu or memory available or too many other sevices already running....

Also...is it a large EE job deployed as a service? ...where karge is hundreds of stagrs?

Ernie

chulett · Post by **chulett** » Sun Jun 12, 2016 4:20 pm

eostic wrote: ...where karge is hundreds of stagrs?

Guess I need to get caught up on the latest techno-babble.

ArndW · Post by **ArndW** » Mon Jun 13, 2016 2:40 am

I have a separate service tier at this site which has a minimal configuration and when I start the Metadata server the Java process consume over 2Gb of memory and uses 100% CPU for a period of about 10 or more minutes.
Next time you start the process, use your favorite monitoring tool (I'm old skool and use "vmstat" and "top") to see if the process is actually using resources or just hanging.

PaulVL · Post by **PaulVL** » Mon Jun 13, 2016 9:09 am

Known issue actually.

There is a patch for it contained in RU 3 I believe.

Something about Websphere cleaning up events.

qt_ky · Post by **qt_ky** » Mon Jun 13, 2016 10:05 am

We also hit that issue. We were having around 20 minute startup times every week. The workaround suggested was to restart more frequently. Instead we just waited for the patch to be released. The patch resolved the problem.