LSM plex detached after disk went faulty

[peter carpenter_1]

I have an old Tru64 system which (unfortunately) is using LSM to create some redundancy. Overnight a disk went bad, but it looks as through the good plex has become detached and stale. I an 100% certain the bad disk is rz8, but the plex containing rz16 subdisks is detached:

root@SPICERS-UK> volprint -Ath
Disk group: rootdg

DG NAME GROUP-ID
DM NAME DEVICE TYPE PRIVLEN PUBLEN PUBPATH
V NAME USETYPE KSTATE STATE LENGTH READPOL PREFPLEX
PL NAME VOLUME KSTATE STATE LENGTH LAYOUT ST-WIDTH MODE
SD NAME PLEX PLOFFS DISKOFFS LENGTH DISK-NAME DEVICE

dg rootdg 1025950414.1025.SPICERS-UK

dm rz10 rz10 sliced 1024 17772484 /dev/rrz10g
dm rz16a rz16a nopriv 0 263455 /dev/rrz16a
dm rz16b rz16b nopriv 0 786505 /dev/rrz16b
dm rz16e rz16e simple 1024 0 /dev/rrz16e
dm rz16g rz16g nopriv 0 3000574 /dev/rrz16g
dm rz17 rz17 sliced 1024 17772484 /dev/rrz17g
dm rz18 rz18 sliced 1024 17772484 /dev/rrz18g
dm rz8a rz8a nopriv 0 263455 /dev/rrz8a
dm rz8b rz8b nopriv 0 786505 /dev/rrz8b
dm rz8e rz8e simple 1024 0 /dev/rrz8e
dm rz8g rz8g nopriv 0 3000574 /dev/rrz8g
dm rz9 rz9 sliced 1024 17772484 /dev/rrz9g

v rootvol root ENABLED ACTIVE 263455 ROUND -
pl rootvol-01 rootvol ENABLED ACTIVE 263455 CONCAT - RW
sd rz8a-01p rootvol-01 0 0 16 rz8a rz8a
sd rz8a-01 rootvol-01 16 16 263439 rz8a rz8a
pl rootvol-02 rootvol DETACHED STALE 263455 CONCAT - WO
sd rz16a-01p rootvol-02 0 0 16 rz16a rz16a
sd rz16a-01 rootvol-02 16 16 263439 rz16a rz16a

v swapvol swap ENABLED ACTIVE 786505 ROUND -
pl swapvol-01 swapvol ENABLED ACTIVE 786505 CONCAT - RW
sd rz8b-01 swapvol-01 0 0 786505 rz8b rz8b
pl swapvol-02 swapvol ENABLED ACTIVE 786505 CONCAT - RW
sd rz16b-01 swapvol-02 0 0 786505 rz16b rz16b

v vol-rz8g fsgen ENABLED ACTIVE 3000574 SELECT -
pl vol-rz8g-01 vol-rz8g ENABLED ACTIVE 3000574 CONCAT - RW
sd rz8g-01 vol-rz8g-01 0 0 3000574 rz8g rz8g
pl vol-rz8g-02 vol-rz8g ENABLED ACTIVE 3000574 CONCAT - RW
sd rz16g-01 vol-rz8g-02 0 0 3000574 rz16g rz16g

v vol01 fsgen ENABLED ACTIVE 17772484 SELECT -
pl plrz10-01 vol01 ENABLED ACTIVE 17772484 CONCAT - RW
sd rz10-01 plrz10-01 0 0 17772484 rz10 rz10
pl plrz18-01 vol01 ENABLED ACTIVE 17772484 CONCAT - RW
sd rz18-01 plrz18-01 0 0 17772484 rz18 rz18

v vol02 fsgen ENABLED ACTIVE 17772484 SELECT -
pl plrz17-01 vol02 ENABLED ACTIVE 17772484 CONCAT - RW
sd rz17-01 plrz17-01 0 0 17772484 rz17 rz17
pl plrz9-01 vol02 ENABLED ACTIVE 17772484 CONCAT - RW
sd rz9-01 plrz9-01 0 0 17772484 rz9 rz9

root@SPICERS-UK> file /dev/rrz8c
/dev/rrz8c: character special (8/16386) SCSI #1 RZ1BB-CS disk #64 (SCSI ID #0) (SCSI LUN #0) errors = 0/26

I have a case raised with HP, and they are advising we try to re-attach the rootvol-02 plex before doing a disk replacement. Which makes sense, but has anyone seen this behaviour before, if so does anyone know how dangerous trying to reattach a good plex to the group containing a known bad disk is?

[Debbie Trenta]

Hi Peter,

rootvol is enabled during the very early boot process. Were there any errors that occured during your boot that would cause rz16a to have become detached? The rest of your rz16 mirrored set (swap and usr) appear to be fine that is why I suspect something occurred during the early boot process.

I agree with the people you have been talking to. Have you tried a # volrecover -sb
I believe that would also work. Once the recovery is complete, you can then replace your rz8 drive.

Whatever you do, backup up your system disk first just to be on the cautious side :-)

Hope this helps.

Debbie