Hi !
I'm having a problem with job daemon taking a lot of memory (HP/UX 11):
ps -efl:
1 S jobdm785 747 746 0 154 20 49891200 32165 2862f34 May 7 ? 166:19 oracle8 (jobdm785:746/PIPE) 71
1 S jobdm785 1199 746 0 154 20 5c755f00 96 2862f34 May 7 ? 0:01 sh_server (jobdm785:746/LSOCKET3
1 S jobdm785 746 742 0 154 20 4cca7e00 31150 2862f34 May 7 ? 98:44 /d1/bse/bin/bshell6.1 723 321 0
1 S jobdm785 723 1 0 154 20 4a5c4f00 288 2862f34 May 7 ? 1:06 ba6.1 ttaad5206m000

Total memory consumption of daemon is around 270MB. Server side oracle process is missing from this list.
Baan is IVc4 SP8. 3-tier setup. Is there any other way than down/up to "clean" it? Seems like it will get mixed up every night.

Porting set:Portingset : 6.1c.06.02
I can't make "bshcmd6.1 -m pid" to work. It gives nothing as result.

I have had a lot of problems with the job daemons. I have found that if you shut them down and start them back up everyday it seems to help a great deal. But you still run the risk of them hanging. Personally I like cron it runs when it supposed to and I have never had a problem with it.

I personally prefer the cron also more than the jobdaemon, also because parallel execution of jobs doesn't work well with jobdaemon.

About the memory, you have to know that a bshell or session can free it's memory however that will never be given back to the OS. So the actual use can be less.
This means that one single session that needs much memory can cause this problem. Probably you can sort that out.

The other thing that can happen is that a session needs a lot of small memory blocks, later on it frees them however because they are small the free blocks can't be used by other sessions.

The 270Mb doesn't need to be in main memory. If it ain't used anymore it's probably paged out.

Conclusion: as long as the system is running fine it doesn't have to be a problem. It will be a problem if it eats all you swap memory.

Regards,
Han