Date: Tue, 22 Jun 2004 13:49:40 +0200 From: Stijn Hoop <stijn@win.tue.nl> To: freebsd-questions@freebsd.org Subject: vinum disks references & crashed state Message-ID: <20040622114940.GI56772@pcwin002.win.tue.nl>
next in thread | raw e-mail | index | archive | help
--aPdhxNJGSeOG9wFI Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Hi, I just had a crash on two of the drives on one vinum RAID-5 volume; fortunately I could recover the data and build a new one. However, while do= ing this I did something 'stupid', which I managed to fix, but I wanted to know what the right way would have been. Like I said one RAID-5 volume was crashed, but I have two volumes on this machine. The other was running just fine, but since I had to swap a lot of drives in/out for recovering the data, I thought I'd disconnect the drives = of the other volume and use those ATA channels to attach my backup drives; all= of this went great, until I had to reattach the drives with the old volume on them. Up until this point, vinum had been showing the still good volume as 'crash= ed' because all subdisks were in state 'referenced'; of course, the correspondi= ng drives could not be found (they were physically detached) so I didn't think this odd. When I reconnected the drives though, it still said the volume was crashed. On the one hand I can understand this; this information was saved on all vi= num drives, so it was also saved on the drives of the other volume. On the other hand, all the drives in the detached volume were consistent, on the same channels etc as before, and they had the right configuration data on them; shouldn't vinum be able to figure out that all disks were still allright and the data intact? I managed to correct this by manually doing 'setstate up <drive>' for all f= our drives in the volume; this worked, and appeared to have no ill effects. Is this the right way to do this? Should I have done something different? Curiously, --Stijn --=20 Light travels faster than sound. That's why some people appear bright until you hear them speak. --aPdhxNJGSeOG9wFI Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (FreeBSD) iD8DBQFA2BzUY3r/tLQmfWcRArLtAJ9NSYP6e2/H7kwhp3Kw8l1tnITtxACeI4f8 APFOvuSqyhHgGCcLhl+4paU= =Fq6P -----END PGP SIGNATURE----- --aPdhxNJGSeOG9wFI--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040622114940.GI56772>