This is an extremely simple script that will return the output of zpool status if the word DEGRADED or FAULTED is seen when running the
zpool status command.
Suggested use is to set this up to run every few hours via cron, which
(when properly
configured) will
email the root user anything printed to stdout. Make sure you have a
filter on your email client that prevents email from your server from
being sent to the spam folder!
You can easily test that this is working by intentionally putting your array in a degraded state. For example:
$ ./check-for-degraded-zfs.sh
# Should return with no output
$ zpool offline POOL_NAME DEVICE_NAME
# Pool is now degraded
$ ./check-for-degraded-zfs.sh
# Should warn and show zpool status
# If you're using cron you'd be well-advised to test that this works
# through cron in addition to directly executing it!
$ zpool online POOL_NAME DEVICE_NAME
# Pool should now recover automatically
Officially, I wouldn't recommend doing this on a pool containing critical data, but you should be able to take a drive offline, see that the script reports an error, and online the drive without major risk (unless you happen to be extremely unlucky and have another drive fail during the very brief window while you're testing this out...)