i have vcs installed on solaris volume manager where root file system is mirrored on svm.
if i like to patch a 2 nodes vcs then from node a i failover the group on node b.
on node a i break down the svm mirror.
initial d0 mirror root is made out of d1 and d2.
metadetach d0 d2
-other way is to break d0 with metadetach d0 d2
edit /etc/vfsab and /etc/system so svm will be disable at reboot.
perform as above.
then stop group on node b,perform as above.
start node a -gabgonfig -c1
start group here
start vcs in node b.
this is more Solaris than VCS. Wrong Forum !? Sounds like the goog old way. Since more than 6 years of Solris 11 its time to think about. If you will stay with Solaris 10, why not delete d2 and use the disk for Live Upgrade and migrate to Zpool/zfs boot ?
It is much easier and less comlicated. Better in an HA environment.
Always cross check Os updates/upgrades with VCS/InfoScale in SORT.