REboot after panic: lv_activatevg

[Coolmar]

We had a system crash with the following error:

Reboot after panic: lv_activatevg: num_pvols

I found lots of posts regarding hardware errors but found nothing on this one. Anyone know what would cause this?

[Robert-Jan Goossens]

Hi Sally,

check this doc.

Title: panic in lv_init_pvols()
Document ID: 8606395040

http://www2.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000078673070

Regards,
Robert-Jan

[Cheryl Griffin · Accepted solution]

It means the LVM headers on your disk are corrupt.

There are several reasons why this could happen, so you need to look at what has transpired on this system.

Did you recently change hardware? restore or edit the LVM configuration? Is this a cloned system?

These messages happened on systems running older versions of the os and are not that common anymore.

[melvyn burnard]

This is generqlly caused when activating VG's at boot time, and their is a corruption or mismatch between the number of disks belonging to the VG in the VGDA and which causes LVM to think that there should be a different
no. of PVs in a volume group 'vg' other than that specified
in the VG header of the VGDA.

I would recommedn logging a call with your local HP REsponse Centre, as fixing this could be very involved.
One question to ask is whether it occurs all the time, or was it just a "one off"

[James R. Ferguson]

Hi Sally:

The LVM metadata is inconsistent in the volume group being activated. IF this is 11.11, patch PHKL_33390 will simply prevent the panic and leave the offending volume group un-activated.

Regards!

...JRF...

[Coolmar]

Thanks for the responses. It was a one time thing and there were SAN problems prior to this. Our OS is 11.11.

[Coolmar]

I also have the following in the syslog:

vmunix: msgcnt 7 vxfs: mesg 003: vx_mapbad - /dev/vg07/lvol1 file system free extent bitmap in au 144 marked bad