Hi,

Is it possible to restore a failed voting disk on CRS 10g R2 without
stopping the cluster?
I have Oracle+CRS 10.2.0.2 on Linux x86, 3 copies of the voting disk on
/dev/raw/raw2,raw12,raw22 (see below). Due to a storage failure raw22
went offline, I have now repaired it and I am trying to restore full CRS
redundancy without shutting down the cluster, but without any success so
far:

# crsctl delete css votedisk
/dev/raw/raw22 Cluster is not in a ready state for online disk removal

# crsctl add css votedisk /dev/raw/raw22
Cluster is not in a ready state for online disk addition

# crsctl query css votedisk
0.   0   /dev/raw/raw2
1.   0   /dev/raw/raw12
2.   0   /dev/raw/raw22 -> THIS IS 'WRONG', CRS and CSSD LOGS SHOW
CLEARLY THAT RAW22 IS OFFLINE

Any comments?

Thanks,
Luca
--
http://www.freelists.org/webpage/oracle-l