mxdomainmgr and cpu usage

[OlivierV]

Hi all.
I have installed SIM 4.2 last week and since the following behaviour occurred three times:
the CPU is used at 100% by sqlserv.exe and mwdomainmgr, sqlmgr.exe start to use all the available memory (from about 100MB in normal usage to more than 500 MB) and all the previously sent eMails are sent again, flooding my mailbox with hundreds of mails. Stopping and restarting these two services does not change anything, I need to restart the server. The time when it starts is the not same and there is no particular task running at these moments. The server is a DL380/2,5 GB of ram, 1.133 MHz running SIM 4.2 for test purposes. There is only Mrtg with SIM on this server, and it helps me to see what happens on this server. Any idea ?
Thanks in advance for any help

[Remon Klein Tank]

Hello Vanoni,

I'm sorry for your e-mail problem, but happy to know that I'm not alone :-)
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=765288

As for the CPU: I had the same problem, and solved it by scheduling a schrink of my SIM database logfile (which kept growing a few gigabytes in size) every 10 minutes. Now my logfile remains small, and the processor only jumps to 100% every now and then.
Too bad our whole system can't be used for pro-active monitoring anymore because of the email bombardments... :-(

Greetings.

[OlivierV]

Hi Remon Klein Tank.
We had the mass mailing issue on IM7 but this was due to a Windows scheduled task : we were stopping SQL each night to free memory but not stopping IM7 service before. When stopping IM7 then SQL and restarting SQL then IM7, all was fine. But it is not the case here.
I agree with the log file, I have set it to be limited to 500 MB and I regularily shrink it when Mrtg graphs show that this size is reached. But I don't see any link between the log file and the mass mailing in our issue. However, I will keep a eye on this.
Thanks for the answer.
Regards.

[OlivierV]

Hi all.
New informations, I don't know if they are related to the CPU usage but it seems there is a relation. Tonight, the CPU spent 1 hour at 100%. It began at 1:00 am, when the daily device identification task runs.
This morning, this task was running again and again and I saw there was 745 instances of this task in the list. This number was increasing continuously. I tried to change the settings of the task (schedule once or disable) but these settings are never applied, the schedule remains the same. If I stop and restart the HP SIM service, the instances are deleted but start to increase again as soon as the service is restarted. It can't modify the settings of this task: I can go to the settings page and change the values, and click on done, but back to the task list, values are not changed. If I stop the instances, they are cancelled and deleted but new ones are created immediately after. The last run time also never changes.
Any idea ?

[OlivierV]

oh, oh, I have found the problem. The sql log file for the SIM db was full. Remon was right about this and the CPU usage. But I wouldn't have thought this could cause such behaviour for the tasks.
This file is limited to 500MB because it can grows very quickly and to a large size.
I will schedule either a logfile shrink and a file size alert on the server to avoid future issues like this. I don't know if it will solve the mass mailing problem we had before. I will keep you aware of the evolution.
See you next year. Have a nice end of year.
Regards.

[Remon Klein Tank]

Glad it worked. If you find out anything more about the mass-mailing problem, please let me know, 'cause I'm running out of options here. I'm now deleting all events (a process that involves a lot of stopping and starting the SIM process because it seems to freeze on big loads).
If that doesn't help I fear that I have to do a complete reinstall...

[OlivierV]

Hi All.
New issue this morning. 8000 mails from the SIM 4.2 server were waiting for me when I arrived at the office. The CPU on the server was at 100 %, stopping and restarting HP SIM and SQL didn't resolve it, I had to restart the server but this didn't help. Another time, the SQL log file was at it's limit (1GB). After shrinking this file and stopping then restarting SIM and SQL, mxdomainmgr was at 100% for 3 minutes then all became normal again. The Log file was quiet for about 2 weeks but started growing suddently yesterday evening. I don't find anything to explain this behaviour, nothing was changed in the SIM settings. The CPU jumped to 100% at 6AM but there was no task at this moment, I suppose this is when the log file was full.
There is only 730 events in the Sim DB (which is under test for the moment, our main CMS is still IM 7.2) so it means that SIM was resending again and again all events.
If I find anything new, I will post it in the forum.