Runaway JDENET_K jobs.

jlowell

Member
Has anyone using v5r3 experienced this? Ocassionally we get one that start using lots of CPU. We check for zombies, but none are present. We try to bring down services to reset the kernel, but even bringing down services does now work as the JDENET_K job won't end. We have to sometimes use ENDJOBABN to get services down. Right after we delete SQLPKG files, CLRIPC, and restart services. This will clear the system; however, we don't always have the luxury of bringing down the system when a JDENET_K job starts to run away. We have to suffer through it as jobs are running that need to finish.
 
We constantly experience similiar issues with runaway kernels, except we never have to go as far to issue an ENDJOBABN. Ending *IMMED always works fine for us. Whenever we see them on our system, 99% of the time there is an F4311EndDoc in the call stack. I recently opened a call with JDE on the issue and they immediatly told us to apply the latest SARs for that function of course, but there's no proof that will fix our issue yet.
 
[ QUOTE ]
We constantly experience similiar issues with runaway kernels, except we never have to go as far to issue an ENDJOBABN. Ending *IMMED always works fine for us. Whenever we see them on our system, 99% of the time there is an F4311EndDoc in the call stack. I recently opened a call with JDE on the issue and they immediatly told us to apply the latest SARs for that function of course, but there's no proof that will fix our issue yet.

[/ QUOTE ]

We are currently up to date on the IBM/Peoplesoft APAR II13878. I don't know about all the current SARS; however, we are in the middle of loadiing 128 performance ESUs, and the lastest baseline ESUs. I don't know how much they will help.
 
Back
Top