stm diagipmid daemon taking up 1 CPU

[Paul F. Carlson]

I have 3 new rp3440's, each running HPUX 11.11 Enterprise Edition. These servers were installed with an Ignite image from an rp7410. The boxes are running very well with one exception: There is an stm process called diagipmid that is taking up 100% of 1 CPU. HP suggested since OnlineDiags was installed from the ignite image, that I swremove it, and reinstall the latest version (A.45.00). I did this, but there is no change. Anyone have any ideas why this process is taking up so much CPU?

Thanks,
Paul

[Paula J Frazer-Campbell]

Paul

Do you have tusc installed? if not download and install from here:-

http://hpux.connect.org.uk/hppd/hpux/Sysadmin/tusc-7.5/

It does not require a reboot.

Once installed attach to the PID and see what is is doing.

tusc

Paula

[Andrew Merritt_2]

Also, what does 'ioscan -fnk' show for the IPMI driver? Is it CLAIMED or UNCLAIMED?

Is the ipmi driver installed in the kernel?


You can safely kill -9 the ipmid process, it won't have an effect on the rest of the OnlineDiags, but obviously it will be worth trying to find the root cause.

[Paul F. Carlson]

ioscan shows:

unknown -1 16 UNCLAIMED UNKNOWN IPMI Controller.

I will also install tusc and take a look.

[Andrew Merritt_2]

I'd predict you'll see something like this from tusc:

access("/var/tmp/APL.SOCK", F_OK) ERR#2 ENOENT
open("/dev/ipmi", O_RDWR|O_NDELAY, 010410) = 1
ioctl(1, 0xc01c6964, 0x68ff1600) ERR#19 ENODEV
close(1) = 0

repeated continuously. I think this indicates a problem with the installation of the IPMI driver in the kernel.

Are you seeing the same problem on all 3 rp3440's, or just on one?

If you can't get ipmi installed, I think the next step is to contact HP support.

[Paul F. Carlson]

Andrew,

After looking in SAM at the driver configuration, the ipmi driver is in an "Out" status. Is there any risk in turning it on and rebooting? I'm not familiar with ipmi, do you know of any documentation on the subject?

Thanks!
Paul

[Andrew Merritt_2]

I'm afraid the kernel and the IPMI drivers are outside my area. I've seen the problem before, but not seen the resolution to it, at least not yet.

[jmb]

I have been seeing this also, when using an Image created from a 7420.

Moving ipmi into the kernel made no difference. kill -9 worked, but that isn't a fix.

Any new info on this problem?

Thanks.