4 stycznia 2012 13:33Hi, is there a way to check why hpcscheduler.exe used 98% cpu ? I had this situation some days ago. It was necessary to restart the service. I have not found useful info in the event viewer -> windows hpc server folder.
4 stycznia 2012 15:56
which exact release are you using ? what is the sizeo fo your cluster ? how many jobs are in the history ?
You can find more events in the event viewer when you enable "Show analytic and debug logs " from the view menu.
12 stycznia 2012 15:29
- which exact release are you using ?
Server version: 3.3.3950.0
Client version: 3.3.3950.0
- what is the size of your cluster ?
We have a cluster of 3 servers (one is head/compute node, the other 2 are compute nodes).
Servers are HP BL685 with HP on board NICs. They are in different subnets 02,03 and 04, connected to 1 Gig switches, running 1 gig full duplex.
The database is local to the same server marked as head node. We use the SQL Express that comes with HPC to host the database.
- how many jobs are in the history ?
how can I find them ? Which api have I to use ? The most recent jobs in the job management are dated Jan 05 so I cannot see the situation when the cpu was 98%