07-27-2011 06:55 PM
We got the issue that the CPU utilization rate goes to 100% and last for about 2 mins when runing supportsave to collect the data on DCX,the currently applied FOS is at 6.3.1b.
Jul 26 17:14:49 raslogd: 2011/07/26-17:14:47, , 9566, SLOT 6 | WWN 10:00:00:05:33:1c:61:00 | FID 128, INFO, dcx_b, The last device change happened at : Tue Jul 26 17:14:44 2011
Jul 26 17:30:17 raslogd: 2011/07/26-17:30:16, , 9568, SLOT 6 | WWN 10:00:00:05:33:1c:61:00 | FID 128, INFO, dcx_b, cpuUsage,is above high boundary(High=75, Low=0). Current value is 100 % at 120 seconds.
Is it normal ?
Another question is how to check if there's any core file existing on switch ? we used "savecore" command before but it no longer work on DCX with 6.x FOS.
Appreciated for any help.
07-28-2011 09:44 AM
DEFECT000264643 describes similar issue and it's fixed in Fabric OS 6.4 branch.
savecore is deprecated since Fabric OS 5.3. Core dump files are saved as part of supportSave routine.
Hope this helps,
07-28-2011 06:48 PM
Thanks for your reply.
I checked the defect and the symptom is described as below:
Whenever they do a supportsave on DCX, they see CUP timeout messages on console. Customer is very sensitive to console messages and would like to know if this is expected behavior or represents a problem. It‘s a single timeout message (IOSo51I) to the cup port happens every time a supportsave is taken.
What's the CUP timeout ? I haven't see the messages in the event log.
Please help to confirm it.
07-29-2011 01:36 AM
CUP is specific to IBM FICON which obviously doesn't apply to your case, but underlying symptoms (high CPU load during supportSave) are similar. That's why I believe updating Fabric OS might help.
Also because 6.3.1b is rather old and is not supported anymore by many major vendors.