Host halted when increasing memory to guest

[Feinberg Parker]

HPUX 11.23 vm v3.0 Has anybody had a problem of the host halting when starting a guest after increasing memory to the guest. We did a hpvmmodify -P host -r 12G
hpvmstop -P Host
hpvmstart -P host

This is when our host halted. Nothing in OLDsyslog or shutdownlog

[Rita C Workman]

No, not so far.

You might try looking at the logfiles at
/var/opt/hpvm/common to see if you can get any info as to what was going on.

Just a thought,
Rita

[Feinberg Parker]

Thanks Rita but the logfiles show nothing unusal.

[Tingli]

Seems your host runs out of memory. Or more precisely, out of swap space.

[Feinberg Parker]

We have 40GB of swap and only 2 guests.

[Rita C Workman]

Well you might start ruling things out.

I might first run swa utility and get any patches for my host server and load them. That will rule out a problem with patches.
Oh....and even before I'd check to make sure all my current patches are in the right (configured) state.

When all else fails, go one step at a time..
Rita

[sujit kumar singh]

Hi

for the swap on the host you should have the swap on the IVM host as Physical RAM+4GB of RAM, that is suggested for IVMs.

apart from that please get the latest patches for the OS and specifically the IVM related patches.
regards
sujit

[Feinberg Parker]

We patched this server just 2 months ago. We really have no info on why it crashed while doing hmvmmodify. Since we seldom make changes to our vm's we are just going to assume it was a freak instance. Thanks all for your ideas.

[Don Morris_1]

Did the whole host OS crash (your original message referencing hpvmmodify/start/stop of "host" (when such commands would only affect Guests) leaves me unclear.

In any event -- did you check /var/adm/crash to see if the crash generated a crashdump? And if so, what do you get from:

cd /var/adm/crash/
kwdb -q4
trace event 0

?

[Rita C Workman]

Gino,

Here's one last thing....do a man on hpvmdevmgmt. You might try using -r option and redirect it to produce a file that may report if there are any problems that could have been the cause of the halt.

...just a thought,
Rita

[Feinberg Parker]

Yes, the host halted. There was nothing in /var/adm/crash or shutdownlog or syslog.log

[Feinberg Parker]

The -r option came out with version 3.5 vm. We are still on 3.0

[Don Morris_1]

If there's no crash either you didn't have enough dump space to store the dump -- or this was a HPMC / MCA type error.

Have you checked the console logs for machine abort errors?