Duplicate PV in VG listing

Question

I have a server that's had a bit of trouble with it's disks (hdisk10/hdisk11/hdisk15 failed, hdisk7 failing) in the past little while but it's happy now. No data loss thankfully as the vanished/stale extents were on an unused part of the filesystem.

Prior to removing the dodgy disks from the volume group:

# lsvg -p udvg
udvg:
PV_NAME           PV STATE          TOTAL PPs   FREE PPs    FREE DISTRIBUTION
hdisk5            active            1082        1080        217..216..216..216..215
hdisk6            active            1082        1080        217..216..216..216..215
hdisk8            active            1082        1082        217..216..216..216..217
hdisk9            active            1082        1082        217..216..216..216..217
hdisk9            missing           1082        1082        217..216..216..216..217
hdisk11           active            1082        1082        217..216..216..216..217
hdisk12           active            1082        1082        217..216..216..216..217
hdisk15           missing           1082        1082        217..216..216..216..217
hdisk14           active            1082        1082        217..216..216..216..217
hdisk13           active            2185        2185        437..437..437..437..437
hdisk26           active            8938        343         00..01..00..00..342
hdisk27           active            8938        343         00..01..00..00..342

Note the duplicate hdisk9 entry. After removing the now-unused disks from the VG:

# lsvg -p udvg
udvg:
PV_NAME           PV STATE          TOTAL PPs   FREE PPs    FREE DISTRIBUTION
hdisk5            active            1082        1080        217..216..216..216..215
hdisk6            active            1082        1080        217..216..216..216..215
hdisk6            missing           1082        1082        217..216..216..216..217
hdisk26           active            8938        343         00..01..00..00..342
hdisk27           active            8938        343         00..01..00..00..342

There's something in the LVM state that's got that extra entry kicking around. Any ideas on how to diagnose and remove the stale entry?

score 1 · Accepted Answer · answered May 12 '14 at 16:34

I got "lucky" - the host rebooted over the weekend due to power problems.

#varyonvg -f udvg
PV Status:      hdisk5  0006c50d44bb48d9        PVACTIVE
                hdisk6  0006c50d44bbc002        PVACTIVE
                hdisk26 000b8f7cd5f2f3da        PVACTIVE
                hdisk27 000b8f7c32afd50a        PVACTIVE
                        0006c50d9930eab5        NONAME

There's the culprit pvid, I was able to remove it:

#reducevg udvg 0006c50d9930eab5
0516-304 putlvodm: Unable to find device id 0006c50d9930eab50000000000000000 in the Device
        Configuration Database.
0516-896 reducevg: Warning, cannot remove physical volume 0006c50d9930eab5 from
        Device Configuration Database.

And everything is happy again.

#lsvg -p udvg
udvg:
PV_NAME           PV STATE          TOTAL PPs   FREE PPs    FREE DISTRIBUTION
hdisk5            active            1082        1080        217..216..216..216..215
hdisk6            active            1082        1080        217..216..216..216..215
hdisk26           active            8938        343         00..01..00..00..342
hdisk27           active            8938        343         00..01..00..00..342

It would be nice to see a way of fixing this that didn't involve an unscheduled reboot, though ;)

Duplicate PV in VG listing

1 Answers1