> On 8/8/05, Steven M. Bellovin <smb <at> cs.columbia.edu> wrote:
> > raidctl is indicating that one of my RAID1 partitions has failed. 
> > What do I do to reconstruct it?
>
> [snip]
>
> > I *think* I say
> >
> >         raidctl -R wd0a radid0
> >
> > I doubt that it's a hardware problem, since the other partition on 
> > that mirrored pair of drives is working properly.

Could be a bad sector though. Take a look at atactl wd0 smart status

> > If it matters, I'm running 2.0.

2.0_STABLE? If so, that improves read reliability (by retrying) and it 
won't fail all components (at boot time, it still seems to be possible 
later on).

> In my experience, what you propose below is the way to go. This is
> assuming I have correctly understood that you think the drive is fine.
> This will start a rebuild of the set, and if the drive is actually
> failing, it should crap out.

For crap out, read panic. :-(

Can we request a pull-up of the fix (in -current and netbsd-3) to 
2.0_STABLE? (see 
http://mail-index.netbsd.org/netbsd-help/2005/08/05/0003.html )

-- 
Stephen