/opt/IBM/InformationServer/Server/PXEngine/bin/osh

[pavan_test]

Hi All,

I am working on data stage 8.1
when I run this command ps -ef | grep bin/osh | wc -l i get around 317

can anyone please tell me what are these files and if i delete them does it improve the performance of etl. (jobs are taking atleast 5min to compile, not sure if both are related)

2637870 1421506 0 20:18:13 - 0:00 /opt/IBM/InformationServer/Server/PXEngine/bin/osh -APT_PMsectionLeaderFlag XXXX 10002 0 30 node1

2666708 1343622 0 20:18:13 - 0:00 /opt/IBM/InformationServer/Server/PXEngine/bin/osh -APT_PMsectionLeaderFlag XXXX 10002 1 30 node2

Thanks
Mark

[chulett]

'ps' does not show files but rather processes, in this case those associated with running PX jobs. And with a large number of processes running, it wouldn't suprise me to hear of other things taking extra time.

[pavan_test]

Thank You. there are 317 process running in this directory.
/PXEngine/bin/osh -APT_PMsectionLeaderFlag XXXX

if i kill the processes in /Server/PXEngine/bin/osh directory will it have any impact on how etl jobs run.

Thanks
Mark

[pavan_test]

I believe these are the defunct processes which gets created while etl jobs are running and usually go away when the job completes.

but i find around317 in
2715724 585894 0 21:04:38 - 0:00 /InformationServer/Server/PXEngine/bin/osh -APT_PMsectionLeaderFlag XXXX 10004

how do I kill them. i fine theis on data stage version 8.1
i haven't seen such behavious in data stage version 7.5.2

Thanks
Mark

[ray.wurlod]

Be very, very careful. Any of these might be a running job. Verify whether each has a PPID (parent process ID) - for a section leader process that would be the conductor process (if on the same machine) or the remote shell that it invoked.