The following is the summarry:-
1) Four disks in server ie (c1t0d0. c1t1d0, c1t2d0, c1t3d0). c1t2d0 is the disk to be replaced.
c1t0d0 and c1t2d0 are mirrors.
c1t1d0 and c1t3d0 are mirrors.
Metadb to be deleted is in c1t2d0s7
a) Mirror d35 has 2 submirrors d38 and d39
d38 is a stripe having 2 members:- stripe 0:- c1t0d0s6
stripe 1:- d36(soft partition), c1t1d0s0
d39 is a stripe having 2 members:- stripe 0:- c1t2d0s6
stripe 1:- d37(soft partition), c1t3d0s0
After deleting metadb and backing up partition table from its mirror,unconfiguring and replacing the disk (c1t2d0), and then configuring it, writing the backed up partition table to it, creating metadb in c1t2d0s7 back.
If i issue metareplace -e d35 c1t3d0s6, would everything be fine? The reason why is asked because c1t2d0s6(failed slice) and a softpartition d37(c1t3d0s0)(its from another disk) forms the submirror d39 which is part of the mirror d39.
Do i have to issue the command metarecover c1t2d0s4 -m -p , or metarecover d36 -m -p ?for re-indexing, please help on this, am really confused.