vPar boot from tape causes vpmon to freak out???

[Eric Yruegas]

I have a SuperDome, 5 cells, 2 I/O chassis. I have one big nPar, with 2 active vPars. We are running vPars 03.03.06.

I've created 4 new vPars, and am preparing to build them out from an Ignite image of one of the running vPars.

I've added the path to the tape drive using the :TAPE suffix, so it shows up properly in the HW list for that specific vPar. When I issue the "vparboot -p {vpar} -B TAPE", the ENTIRE nPar reset. Yikes! I suspect the vpar monitor had a cornary for some reason and rebooted. The tape hadn't even started to read when it happened.

I've done this in the past on rp8420's, so I do know that the functionality works - but I'm now a bit gun-shy to try this again. I'm going to open a case with HP, but wanted to see if anyone else had seen this behavior before.

Worst case, I can boot the whole thing as an nPar and Ignite to the target disks, but I'm trying to avoid downtime on the 2 active partitions.

Thanks all!

[Duncan Edmonstone]

Eric,

Just to double check - this is a *local* tape drive, and not a *SAN* based tape drive yes?

HTH

Duncan

[Eric Yruegas]

Yup - locally attached tape.

And to followup - I opened a case with HP, and spent a good number of hours on the phone with them. Apparently whatever is happening is at such a low level that no logging feature is really capturing the partition reset.

We'll be looking at it again tomorrow, but all engineers so far have agreed it's the strangest thing they've seen so far. Without a log or dump or real tombstone, its very difficult to trace.

I'll post more when I know more... (in case anyone's interested)

[Ramones]

Hi Eric

Can you post parstatus and vparstatus outputs?


Regards,
STP