Veritas Volume Manager 5.0.1 Troubleshooting Guide, HP-UX 11i v3, First Edition, November 2009

Recovering the root volume after VxVM emergency startup

To recover the root volume after VxVM emergency startup

After you have used vx_emerg_start to start VxVM, use the vxprint command

to determine the configuration state.

One common problem is that all the plexes of the root volume, rootvol, are

stale. This would be shown in the STATE field of the vxprint output as follows:

TYNAME ASSOC KSTATE LENGTH PLOFFS STATE ...

vrootvol root DISABLED 393216 - ACTIVE ...

plrootvol-01 rootvol DISABLED 393216 - STALE ...

sdrootdisk01-02rootvol-01 ENABLED 393216 0 - ...

plrootvol-02 rootvol DISABLED 393216 - STALE ...

sdrootdisk02-02rootvol-02 ENABLED 393216 0 - ...

(The TUTIL0 and PUTIL0 fields have been removed from the vxprint output

for readability.)

The root volume can usually be repaired by using the following command:

# vxvol -g bootdg -f start rootvol

If the root volume is mirrored, recovery is started. Wait until recovery

completes and the command exits. Then run the fsck command and mount

the root file system as shown here:

# fsck -F vxfs -o full /dev/vx/rdsk/bootdg/rootvol

# mkdir /tmp_mnt

# mount -F vxfs /dev/vx/dsk/bootdg/rootvol /tmp_mnt

The following form of the vx_emerg_start command combines all these

operations in a single command to recover the root volume and its mirrors,

check the root file system, and mount it:

# vx_emerg_start -m hostname

When you have recovered the volumes on the VxVM root disk, and performed

any other necessary repairs, reboot the system:

# reboot

Recovering from boot disk failure

Recovering a system by booting from recovery media