Monitoring almost-non-fail
I almost got it right. When I installed my home server, I was sure to set up mdadm monitoring so it would email me if a drive in my RAID array failed.
I idly checked /proc/mdstat today and saw that my array is degraded. I checked logs and discovered that this happened in July.
Now, mdadm did send mail, though I didn’t receive it. If I had tested outgoing email to that account, I would have noticed a delivery failure from my own blacklisting of dynamic IPs.
Oops.
Eugene Crosser November 26, 2012 01:59
Quis custodiet ipsos custodes?
I use xmpp for notifications.
If you do that, you lose notifications because after python upgrade last year, xmppsend stopped working ;)
Derrick Devine November 26, 2012 09:35
Had the same thing happen to me. I setup things to use a gmail account to send it next time.
Michael K Johnson November 26, 2012 10:44
I’m getting three (larger) drives to reconstitute the array, so that I have a hot spare.
I’ll probably fix notifications, too.
Imported from Google+ — content and formatting may not be reliable