what is the status of this vg?

[stephen peng]

guys,
check attached file. I've never met this situation of the status of a vg, what could lead to this situation?

regards
Stephen

[Torsten.]

This looks strange, try

vgchange -a y vg02

and do the vgdisplay again.

[stephen peng]

hi Torsten,
it says:
RX6600_4[/]vgchange -a y vg02
Volume group "/dev/vg02" is already active in requested mode.
confused me again.and vgdisplay remained the same as before.

regards
Stephen

[stephen peng]

and
RX6600_4[/]vgchange -a n vg02
vgchange: Couldn't deactivate volume group "vg02":
Device busy

[Torsten.]

What is the result of

# lvmadm -l -V 1.0

(-l is lower case "L")

[stephen peng]

RX6600_4[/]lvmadm -l -V 1.0
--- Version 1.0 volume groups ---
VG Name /dev/vg00
PV Name /dev/disk/disk1_p2

VG Name /dev/vg02
PV Name /dev/disk/disk34
/dev/disk/disk39

[stephen peng]

it remained me that I met a case once, would someone re-created vdisk that presented to this server at EVA site that caused this problem?

[stephen peng]

http://h30499.www3.hp.com/t5/System-Administration/unknown-reason-cause-VG-activation-failed/m-p/4754909#M388567

I am somewhat clear what happened, but I don't think vgdisplay show correct status of the vg in this situation, vg and lvs corrupted, but lvdisplay shows the lvs' status well.

regards
Stephen

[Torsten.]

This could be true.

What are your SAN folks doing?

Could you try "pvdisplay" against the disks?

[P Arumugavel]

hi,

Please read the below threads for same case.

http://h30499.www3.hp.com/t5/LVM-and-VxVM/LV-Status-available-syncd-but-PV-unavailable/m-p/5119404#M46152

http://h30499.www3.hp.com/t5/System-Administration/PV-Status-Unavailable/m-p/5264015#M472521

Please update with your observation and results.

Rgds...

[Torsten. · Accepted solution]

IMHO these threads are not related.

The problem is, the VG status is available, but both disks are unavailable, so the LV Status can't be available/syncd.


If the vdisks on the EVA are really re-created, the disks are no longer PVs, so the pvdisplay command will fail.

[stephen peng]

guys,
please check attached file for pvdisplay -v output.
1,only PV Status is not "well", and other details seemed right. It is beyond my usual understanding about HP_UX; why it could still show LVs details? where is the LVs' or PVs' information stored? and pvdisplay did not fail,right?
2,when vg02 was in activated status, could vdisk at EVA site be re-created without prompt?


regards
Stephen

[Hakki Aydin Ucar]

>Stephen: vgchange: Couldn't deactivate volume group "vg02":

any mounted file system on it ?

# bdf
# fuser -u /dev/vg02/mylvol1
# fuser -u /dev/vg02/lvol2

According to pvdisplay output , you can try;

# pvchange -a y /dev/disk/disk34