mdadm first harddrive fail does not result in proper remount of fs

jalexxi · 02-23-2009, 02:07 PM

I've been setting up a server with software RAID 1, of which the function is to be both a backup and to keep the server from ceasing it's normal operation (as both a fileserver and a printserver) when one of the drives fails. The problem is, if I make the second drive in the array md0 (hdc, secondary master) fail, it remounts the fs properly and the system resumes, but if I do the same to hda (primary master), it crashes and never recovers. Am I doing something wrong, and what can I do to prevent this?

Thanks in advance!

mostlyharmless · 02-23-2009, 02:26 PM

Got a couple of questions to make the situation clearer to all:

How are you making the drives fail?
(a) marking them as a failed drive using mdadm
(b) something else

How is the system reacting?
(a) nothing happens until I reboot, then I get the following message
(b) it immediately crashes, then I get the following message
(c) something else

jalexxi · 02-24-2009, 07:45 AM

I've found out my method of making it fail (yanking out the power) was too catastrophic under most circumstances, and it was just a coincidende it happened with the second HD every time. I've now gotten a failure with the first one as well.
This problem does not occur when marking it as failed in mdadm... So that solves that problem, sort of. If it doesn't fail too catastrophically, it recovers. I can live with that, I think. Thanks for the response anyway, mostlyharmless!