Solstice Disksuite: Recovering the quorum after a disk failure on the system start-up

In a scenario where we lost one disk during the boot, the system start-up couldn't be finalised due to loss of the quorum. So the system will prompt to enter in single user Enter in single user and type the following.

# metadb -i
This will list all meta databases. The replicas that have problems would be flagged with “M”.
So, would be necessary to remove all the replicas from the meta database in order to get the correct
quorum for the mirror to work.
Example
flags first blk block count
M 16 1034 /dev/dsk/c1t0d0s7 <-- failed disk0 M 1050 1034 /dev/dsk/c1t0d0s7 <-- failed disk0 a m p luo 16 16 /dev/dsk/c1t1d0s7 a p luo 1050 1034 /dev/dsk/c1t1d0s7 # metadb -d /dev/dsk/c1t0d0s7 Ignore the messages saying cannot change the file “read-only file-system” # reboot The system will boot from the mirrored disk and this time without quorum issues. Reconnecting the disk0 after bootable mirror test Assuming that you are doing a full test, and the previous Troubleshooting procedure was followed, then issue the following commands: # metadb -a -f -c2 /dev/dsk/c1t0d0s7 The above command will recreate the meta database on the disk0 and once that done, and since it was not a disk failure/corruption. So, the meta structure remains in the disk, the SDS immediately starts the synchronization in all file-systems. # metastat or # metastat grep “%” Disabling the Quorum Rule # echo “set md:mirrored_root_flag=1” >> /etc/system