Puzzling Case: HP-UX would just stop, rest at BCH, no Crash Dump,etc.

[Alzhy]

We've an rp8420 that all of a sudden started acting weird. In the span of 24 hours, it simply "stopped" - no indications it crashed or panic'd. The system simply went down to BCH mode.

We've raised a tkt with HP but seems to be getting no to slow response in the absence of dumps or indications it crashed.

So I tried scanning FPL/SEL logs in the SEL logs, I found the following that matches exactly the time the environments on said rp8420 was reported to be unavailable:

8408 SFW 0,3,1 *3 648016c531e00947 0f0303ffffffff8d IO_LINK_FATAL
8408 05/14/2007 13:33:38
8409 SFW 0,3,1 *3 648016c531e00949 0f0303ffffffff8d IO_LINK_FATAL
8409 05/14/2007 13:33:38
8410 SFW 0,3,1 *3 648016c531e0094b 0f0303ffffffff8d IO_LINK_FATAL
8410 05/14/2007 13:33:38
8411 SFW 0,2,0 *3 648016c520e0094d 0f0303ffffffff8d IO_LINK_FATAL
8411 05/14/2007 13:33:38
8412 SFW 0,2,0 *3 648016c520e0094f 0f0303ffffffff8d IO_LINK_FATAL
8412 05/14/2007 13:33:38
8413 SFW 0,2,0 *3 648016c520e00951 0f0303ffffffff8d IO_LINK_FATAL
8413 05/14/2007 13:33:38
8414 SFW 0,2,0 *3 648016c520e00953 0f0303ffffffff8d IO_LINK_FATAL
8414 05/14/2007 13:33:38
8415 SFW 0,3,1 *3 648016c531e00955 0f0303ffffffff8d IO_LINK_FATAL
8415 05/14/2007 13:33:38

So - do you think this is a HW issue?

BTW, the rp8420 is carved as a single nPar with 4 vPars. It's not set to auto start and will always rest at BCH on reboot/restart of the frame.

Again, what is puzzling is Console Log show no indication of a crash / panic or any other messages. It simply dropped to the BCH prompt.

[Alzhy]

FYI!

It turns out and HP concluded - we have a BAD cell board that needs to be replaced.

[John Payne_2]

I would have guessed a problem with the IO card cage (you could have seen that by looking for absense of all lights on the card cage) , but it's good they figured it out.

John

[Denver Osborn · Accepted solution]

We had the same thing happen on an rp7420. Box rebooted w/out any panic. From the server logs at the it showed the cell board lost power. No indication from the host side that there was a problem other than the unexplained reboot.

-denver