Multiple task instances in pending state

[OlivierV]

Hello.
I have three tasks that have multiple instances in a pending state. One is the hw status polling task for servers (schedule interval 15 minutes, task duration 7 minutes average, 5 to 8 pending), one is the hw status polling task for non server (schedule interval 1 hour, task duration 5 minutes average, 1 or two pending) and the third one is a custom hw status polling task for about 30 servers that need to be polled more often (schedule interval 5 minutes, task duration 3 minutes average, 10 to 18 pending).
All these tasks are ran at the CMS start (restart the CMS service every night at 9 PM), but deselecting this option does not change anything. The number of pending instances is decreasing along the day.
Is this situation normal or not (In the past, I saw days with only one complete instance for each task), and what should I do to correct that ?
The CMS is a ML370G2, 2xPIII 1.133 GHz, 2.5 Gb of ram and 2x100Mb/s nics in teaming. About 400 servers and 250 other devices, mainly iLO/ Riloe... OS W2003 std and SQL 2000 SP3a
Thank you.

[OlivierV]

Hello.
Since yesterday, these tasks are quiet, no more pending ones. Could it be due to one task having some problems and making the other ones having many instances. We deleted some servers in the past days, can one machine causing some troubles ?

[OlivierV]

Back to this thread.
I had to restart the HP SIM service during the day because of a login issue.
After the restart, I checked the HW status polling for servers task. There was only one instance.
After about 1 minutes, a second instance, in pending state, appeared, before the previous one was finished.
The HW status polling task is scheduled every 15 minutes and takes less than 7 minutes to complete.
Can someone explain me me this behaviour ?
Many thanks.

[HP Systems Insight]

I have the exact same issue (SIM 5.0 SP4 on W2K3) although there are periods of time where SIM will be running just fine without these multiple instances of polling tasks queueing up out of control. I think it may have something to do with whether or not there are an "critical" systems at the time but not 100% sure about this. We're running IP only - no IPX.

Any help would be greatly appreciated.

MJ

[OlivierV]

You are right Michael.

I checked the tasks this morning where there was no 'critical' device and it all goes well.
It looks like SIM adds a instance each time it finds a device that is not responding, leading to tenths of instances. However, this does not seem to have bad effects on the SIM server itself (CPU and memory usage)

Can the HP guys give us some explanations about this behaviour ?

Regards.

[OlivierV]

This morning, 7 devices are Critical. But there are no multiple instances.
I do not understand that mess.

[OlivierV]

The cause is known (devices in critical state) but no answer from HP