[Wolves] Software RAID reconstruction

[Peter Evans]

On Tuesday 26 August 2008 10:46:54 Adam Sweet wrote:
> Hi everyone

I'm only passingly familiar with the whole subject, but:

You've got two discs in a mirror c/d?

My understanding - mirror raid is resistant to a single drive failure.

If d failed, c should take over a d would be failed out of the array, or 
marked bad, whatever.

If your machine locked up when d failed isn't it possible that c may have also 
failed, and possibly some time ago?  This would explain why the raid is 
complaining it's not 'fresh'.

If you're getting repeated I/O errors from 'd' can't you manually mark the 
block as bad, and see how far you get?

Since 'c' isn't playing nice - you might try seeing what's actually on it?  If 
it's being the synch from d isn't working or has been attempted lots of times 
this probably won't yield much though.

As I said, it's a long time since I've worked on anything like this, and even 
then it was only high-level stuff.  I also appreciate it's a live box and the 
constraints that puts on you trying stuff, but given c appears to be useless 
atm taking it out for a look-see might be possible?

Wish you luck.
-- 
Regards,

Pete Evans