How to restore a disk marked failed caused by link failure or I/O transport failure, but not because of the bad disk
Article ID: 100006454
Updated On:
Resolution
If vxdisk marked failed, which is caused by the link offline, I/O transport failure, but not because of the bad disk, the following steps can help to recover the volume:
1. Restore the VERITAS Volume Manager (tm) information to the failed disk. From the
2. Use the vxdiskadm utility to remove and replace the failed drive:
Choose item no. 4 to 'Remove a disk for replacement'
Note: The data in the public region will not be lost.
1. Restore the VERITAS Volume Manager (tm) information to the failed disk. From the
vxdisk list output: # vxdisk list - - testdg01 testdg failed was:c1t12d0s2 - - testdg02 testdg failed was:c1t13d0s2 2. Use the vxdiskadm utility to remove and replace the failed drive:
# vxdiskadm Choose item no. 4 to 'Remove a disk for replacement'
Note: The data in the public region will not be lost.
Before use the disk as a replace disk, initialize it first in another session by:
vxdisksetup -if accessname <<<<<< if format/attribute is not default, need use same attribute, format as previous.
Note: Reinitialize the disk only to restore Volume Manager information to the private region.
Then choose item no. 5 'Replace a failed or removed disk'
After running items 4 and 5, the
3. Recover the volume by performing the following:
Note: This changes the plex state from RECOVER to STALE
After running items 4 and 5, the
vxdisk list output should be similar to the following output: # vxdisk list c1t12d0s2 sliced testdg01 testdg online c1t13d0s2 sliced testdg02 testdg online c1t14d0s2 sliced testdg03 testdg online 3. Recover the volume by performing the following:
# vxprint -htg testdg dm testdg01 c1t12d0s2 sliced 2179 8920560 dm testdg02 c1t13d0s2 sliced 2179 8920560 dm testdg03 c1t14d0s2 sliced 2179 8920560 v testvol - DISABLED ACTIVE 204800 RAID - pl testvol-01 testvol DISABLED RECOVER 204864 CONCAT - RW sd testdg01-01 testvol-01 testdg01 0 102460 c1t12d0 ENA sd testdg02-01 testvol-01 testdg02 0 102460 c1t13d0 ENA sd testdg03-01 testvol-01 testdg03 0 102460 c1t14d0 ENA # vxmend -g testdg -o force off testvol-01 # vxmend -g testdg on testvol-01 Note: This changes the plex state from RECOVER to STALE
# vxprint -htg testdg dm testdg01 c1t12d0s2 sliced 2179 8920560 dm testdg02 c1t13d0s2 sliced 2179 8920560 dm testdg03 c1t14d0s2 sliced 2179 8920560 v testvol - DISABLED ACTIVE 204800 RAID - pl testvol-01 testvol DISABLED STALE 204864 CONCAT - RW sd testdg01-01 testvol-01 testdg01 0 102460 c1t12d0 ENA sd testdg02-01 testvol-01 testdg02 0 102460 c1t13d0 ENA sd testdg03-01 testvol-01 testdg03 0 102460 c1t14d0 ENA # vxmend -g testdg fix clean testvol-01 Note: This makes the plex in CLEAN state
# vxprint -htg testdg dm testdg01 c1t12d0s2 sliced 2179 8920560 dm testdg02 c1t13d0s2 sliced 2179 8920560 dm testdg03 c1t14d0s2 sliced 2179 8920560 v testvol - DISABLED ACTIVE 204800 RAID - pl testvol-01 testvol DISABLED CLEAN 204864 CONCAT - RW sd testdg01-01 testvol-01 testdg01 0 102460 c1t12d0 ENA sd testdg02-01 testvol-01 testdg02 0 102460 c1t13d0 ENA sd testdg03-01 testvol-01 testdg03 0 102460 c1t14d0 ENA # vxvol -g testdg start testvol # vxprint -htg testdg dm testdg01 c1t12d0s2 sliced 2179 8920560 dm testdg02 c1t13d0s2 sliced 2179 8920560 dm testdg03 c1t14d0s2 sliced 2179 8920560 v testvol - ENABLED ACTIVE 204800 RAID - pl testvol-01 testvol ENABLED ACTIVE 204864 CONCAT - RW sd testdg01-01 testvol-01 testdg01 0 102460 c1t12d0 ENA sd testdg02-01 testvol-01 testdg02 0 102460 c1t13d0 ENA sd testdg03-01 testvol-01 testdg03 0 102460 c1t14d0 ENA 4. Run a file system utility check and mount the file system. Below is the command for 'vxfs' file system:
# fsck -F vxfs /dev/vx/rdsk/testdg/testvol Note: The file system is clean - log replay is not required
# mount -F vxfs /dev/vx/dsk/testdg/testvol /testvol # df -k /testvol File system kbytes used avail capacity mounted on /dev/vx/dsk/testdg/testvol 44429208 39734216 4658376 90% /testvol Issue/Introduction
How to restore a disk marked failed caused by link failure or I/O transport failure, but not because of the bad disk
Was this article helpful?
Yes
No
Powered by
