Disk Failure

[Carlos Vanegas_2]

I have a volume vg01 with 3 disk, like:
vg01
c1t1d0
c2t1d0
c3t1d0

The disk c1t1d0 failure, I want to know the correct procedure to change it.

Thanks

[Bharat Katkar]

Hi,
Do you have lv's c1t1d0 mirrored/stripped across other two disks??
If it's plain disk without any mirroring / stripping then ...
1. vgreduce c1t1d0
2. remove the failed disk and insert new
3. pvcreate c1t1d0
4. vgextend with c1t1d0
5. lvcreate/newfs on c1t1d0

That's all.
regards,

[Carlos Vanegas_2]

Do you have lv's c1t1d0 mirrored/stripped across other two disks??
No, I dont have mirror

But, I need to create lvol and file system again? I lost the file system?

5. lvcreate/newfs on c1t1d0
But

[Steven E. Protter]

if part of the filesystem exists on another disk, you may be able to extendfs.

otherwise its rebuild newfs and restore the data from other sources.

SEP

[Carlos Vanegas_2]

Steven,
Yes I have lvol1 in that vg01, with 3 disk and mounted in /carlos file system, my question is, if I can change the disk without losing the /carlos file system.

thanks

[Patrick Wallek]

If the disk wasn't mirrored and the LVOL1 existed only on disk c1t1d0 and disk c1t1d0 went bad, then yes you lost the filesystem AND the data.

I hope you had the data backed up somewhere.

[Torsten.]

Hi Carlos,

replace the disk, restore LVM information and activate the VG.

vgcfgrestore -n vg01 /dev/rdsk/c1t1d0
vgchange -a y vg01

display the information about lvols on that disk

pvdisplay -v /dev/dsk/c1t1d0

(only the first screen is useful here)
look for "Distribution of physical volume" information.

create a new filesystem on that volume(s), mount and restore your backup.

[Carlos Vanegas_2]

Steve: I dont have a mirror and lvol1 is in the 3 disk:

--- Logical volumes ---
LV Name /dev/vg01/lvol1
VG Name /dev/vg01
LV Permission read/write
LV Status available/syncd
Mirror copies 0

--- Distribution of logical volume ---
PV Name LE on PV PE on PV
/dev/dsk/c1t1d0 4340 4340
/dev/dsk/c2t1d0 4340 4340
/dev/dsk/c3t1d0 1320 1320

but,I have a backup

[Raj D.]

Hi Carlos ,

Can you provide the output of

Well , you can do the following:

1. # vgreduce /dev/vg01 /dev/dsk/c1t1d0
2. Connect the new disk , assuming its hot swap disk.
3. # vgcfgrestore -n /dev/vg01 /dev/rdsk/c1t1d0
4. # vgchange -a y /dev/vg01
5. Restore the backup from tape.

Cheers,
Raj.

[Torsten. · Accepted solution]

@ Raj D.

If you do a vgreduce, you'll have to do a vgextend instead of vgcfgrestore, because the conf information is gone, right?

[Raj D.]

Torsten you are right, its my mistake ,

Carlos , You can do vgcfgrestore , without vgreduce , that will work.

hth ,
Raj.

[Devender Khatana]

Hi,

As your file system is spread across this disk, replacing it will make file system unavailable. Allthough if your other two disks have enough free space in them then these extents will be moved to the other two disks while doing a vgreduce.

A complete output of "vgdisplay -v /dev/vg01" will tell all that. Can you attach the same?

Not having a free space in other two disks means you first have to attach a new disk in addition to failed disk and will have to do vgextend to new disk and then a vgreduce from the failed disk. In this all extents of the failed disks will be moved to new disk.

HTH,
Devender