Hi. We have a hard drive failing in one of our production servers.
The server has two drives, mirrored. It is split between UFS with SVM, and ZFS.
Both drives are setup as follows. The drives are c0t0d0 and c0t1d0. c0t1d0 is
the failing drive.
slice 0 - 3.00GB UFS (root partition)
slice 1 - 1.00GB swap
slice 3 - 4.00GB UFS (var partition)
slice 4 - 60GB ZFS (mirrored slice in our zfs pool)
slice 6 - 54MB metadb
slice 7 - 54MB metadb
I think I have the plan to replace the harddrive without interrupting either
the SVM mirrors on slices 0,1,3 or the ZFS pool which is mirrored on slice 4.
I am hoping someone can take a quick look and let me know if I missed anything:
1) Detach the SVM mirrors on the failing drive
===========================================
metadetach -f d0 d20
metaclear d20
metadetach -f d1 d21
metaclear d21
metadetach -f d3 d23
metaclear d23
2) Remove the metadb's from the failing drive:
===========================================
metadb -f -d c0t1d0s6
metadb -f -d c0t1d0s7
3) Offline the ZFS mirror slice
===========================================
zpool offline <poolname> c0t1d0s0
4) At this point it should be safe to remove the drive. All SVM mirrors are
detached, the metadb's on the failed drive are deleted, and the ZFS slice is
offline.
5) Insert and partition the new drive so it's partitions are the same as the
working drive.
6) Create the SVM partitions and attach them
===========================================
metainit d20 1 1 c0t1d0s0
metattach d0 d20
metainit d21 1 1 c0t1d0s1
metattach d1 d21
metainit d23 1 1 c0t1d0s3
metattach d3 d23
7) Add the metadb's back to the new drive
===========================================
metadb -a -f -c2 c0t1d0s6 c0t1d0s7
8) Add the ZFS slice back to the zfs pool as part of the mirrored pool
===========================================
zpool replace hrlpool c0t1d0s4
zpool online c0t1d0s4
DONE
The drive should be functioning at this point.
Does this look correct? Have I missed anything obvious?
I know this isn't totally ZFS related, but I wasn't sure where to put it since
it has both SVM and ZFS mirrored slices.
Thanks in advance for any input.
This message posted from opensolaris.org
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
