One of the cool and often overlooked features in Veritas Volume Manager is the failure notification mechanism. This facility provides automated notifications when problems are detected with Veritas managed disks, plexes, subdisks and volumes. These notifications are active by default, and will generate an e-mail to the user root each time a failure is detected. These e-mail notifications take the following form:
To: root@tigger
Subject: Volume Manager failures on host tigger
Content-Length: 240
Failures have been detected by the VERITAS Volume Manager:
failed disks:
c1t6d0
failed plexes:
raid5vol-P08
Since e-mails will be sent to the root user by default, it is often beneficial to create a root alias in the /etc/aliases file:
$ grep "^root" /etc/aliases /etc/aliases.db
root: admins@prefetch.net
If you would like messages to be sent to a user other than root, you can add additional users to the line that starts vxrelocd in the vxvm-recover init script:
$ grep vxrelocd /etc/init.d/S95vxvm-recover
vxrelocd root alerts &
I have been using this facility for years to get advanced notifications, which has helped me avoid disaster on more than one occassion.