In article <0001HW.BE22C8EA00096080F03055B0@news.diapason.com >,
Georges Tomazi <gt@diapason.com> writes:
>
> Hi -
>
> Last night I had a problem with a Solaris 9 server. A 67 Gb slice (on a 73 Gb
> RAID-1 disk) failed and the slice was taken over by a hot spare. This
> morning, metastat was showing the following :
>
> # metastat d3
....
> d33: Submirror of d3
....
> Device Start Block Dbase State Reloc Hot Spare
> c0t2d0s3 0 No Okay Yes c0t4d0s7
....

> However, the disk seems to be OK (less than a year old). No bad blocks and
> the others slices on it are just fine.

So why did disksuite fail it? It must have had some reason for doing so
that is likely to be logged somewhere. Disksuite is pretty aggressive
at failing devices (it dosn't allow many errors before chucking it out
completely) but I've always seen at least one error that explains it.

I usuall (in the cases where the disk is responsive at all) do a
format/analyze/read/repair and try replacing it just to see if it was a
single bad block. Sometimes works.

> So before physically replacing it, I'd
> like to put back the submirror on it and see how it goes. If it fails again
> I'll replace it.
>
> I'm just not 100% sure how to enable again the mirror. It would be great if
> someone could confirm if that's the right way to do it:
>
> # metadetach d3 d33 (detaching the failed submirror)
> # metaclear d33 (clearing it)
> # metainit d33 1 1 c0t2d0s3 -h hsp000 (creating the submirror and allocating
> again the hot spare pool)
> # metattach d3 d33 (attaching and synchronizing the submirror)

Seems complicated. How about just

metareplace -e d3 c0t2d0s3

> I'm not sure about the third step. In Sun's documentation, they suggest a
> command like "metainit d33 2 1 c0t2d0s3 1 c0t3d0s3".

That's a concatenation. Don't think you want that...

> I'm concerned about the risk of loosing the data on the working submirror
> (c0t3d0s3)...
>
> After enabling again the submirror, the hot spare should be automatically
> released, right ?

If the metareplace succeeds. And if the metareplace fails, the hot
spare should still be in place.

--
-Peter Tribble
MRC Rosalind Franklin Centre for Genomics Research
http://www.rfcgr.mrc.ac.uk/~ptribble/ - http://ptribble.blogspot.com/