Oracle clusterware locking the disk devices

Dec 6, 2016 12:10PM

ARCHITECTURE

Version: Oracle Clusterware 11.1.0.7

Architecture: 2-node Oracle RAC 11.1.0.7 + ASM + CRS on AIX 6.1.0.0

HISTORY

We successfully moved 2 voting disks from one SAN to another

by adding 2 new disks then deleting the 2 old ones (backup + crsctl add + crsctl delete)

PROBLEM

Now we cannot detach physically the old disks because the ocssd.bin process is locking the disk devices on both nodes:

##Node 1

$ lsof|grep -i -E 'rvotgln|rvotmgh'

ocssd.bin 8454166 oracle 16u VCHR 21,28 0t0 788 /dev/rvotmgh

ocssd.bin 8454166 oracle 18u VCHR 21,27 0t0 787 /dev/rvotgln

##Node 2

$ lsof|grep -i -E 'rvotgln|rvotmgh'

ocssd.bin 11206854 oracle 16u VCHR 20,44 0t0 33418 /dev/rvotgln

ocssd.bin 11206854 oracle 20u VCHR 20,45 0t0 33420 /dev/rvotmgh

Database - RAC/Scalability (MOSC)