Cluster

• Cluster maintenance update (see “Storage in HA: manutenzione ordinaria e straordinaria.”, RT N.41, febbraio 2012, Ceris-CNR(PDF))

* SystemRescue 3.1.1 changes disk naming from /dev/cciss/c0dx to /dev/sdy (e.g. /dev/cciss/c0d0p2 → /dev/sda2).
Node backup:

root@uclu2:~# crm node standby uclu2
root@uclu2:~# reboot [insert SystemRescue USB]

From SystemRescue:

# mount /dev/sdc2 /mnt/backup
# fsarchiver -vo savefs /mnt/backup/uclu2.fsa /dev/sda1
# umount /mnt/backup
# reboot [remove SystemRescue USB]

* After DRBD package update:

It is very important to remove startup script BEFORE reboot to avoid cluster re-sync and/or spli-brain:

# update-rc.d -f drbd remove