[sllug-members]: Replacing a failed HD in a software mirrored array

[Marc Christensen]

Knight Walker wrote:

> 
> The drive may not actually be failed.  I've seen quite a few situations
> where the RAID will kick out a drive for a transitory issue.  For me,
> removing the drive from of the RAID then re-adding it fixes the problem.
> If the problem reoccurs again, then it's time to replace the drive, and
> it's about the same procedure except for shutting down the machine to
> swap the physical part (If you don't have hot-swap).

So, I am also seeing failed reads in the /var/log/messages so I do think 
the drive is toast.

The funny er....sad thing is that looking at the output from 'mdadm 
--examine /dev/hde1', it shows that the last update was on Mon Oct  9 
12:30:19 2006!  So, I've been running on one drive since October. 
That's nice.

I actually have a script that runs and checks the array for errors. 
Just not in a crontab so that will be fixed once I have the new drive in 
place!

--
Marc Christensen
http://blog.mecworks.com