From mboxrd@z Thu Jan 1 00:00:00 1970 From: "David M. Strang" Subject: Re: raid5: Disk failure on sdm, disabling device Date: Thu, 1 Sep 2005 03:38:01 -0400 Message-ID: <00e001c5aec8$151f1d20$c100a8c0@NCNF5131FTH> References: <003401c5ae61$d982b5d0$c100a8c0@NCNF5131FTH> <4316050B.1060805@tls.msk.ru> <005201c5ae63$ef23e6f0$c100a8c0@NCNF5131FTH> <43160DF2.6000504@tls.msk.ru> <007401c5ae6a$7e9cd160$c100a8c0@NCNF5131FTH> <1125520778.3194.18.camel@x1-6-00-0d-60-fd-e8-4c> <008b01c5ae6c$e4db5e90$c100a8c0@NCNF5131FTH> <1125521616.3194.22.camel@x1-6-00-0d-60-fd-e8-4c> <00ab01c5ae6f$1bb27410$c100a8c0@NCNF5131FTH> <43162122.703@tls.msk.ru> <00c901c5ae73$b80e3ca0$c100a8c0@NCNF5131FTH> <1125526791.3194.32.camel@x1-6-00-0d-60-fd-e8-4c> Mime-Version: 1.0 Content-Type: text/plain; format=flowed; charset="iso-8859-1"; reply-type=original Content-Transfer-Encoding: 7bit Return-path: Sender: linux-raid-owner@vger.kernel.org To: Forrest Taylor Cc: Michael Tokarev , Linux RAID List-Id: linux-raid.ids Forrest Taylor wrote: > > Michael Tokarev wrote: > > > David M. Strang wrote: > > > [] > > > > > > > Is there something a little deeper to this error message? > > > > > > > > Aug 31 04:48:15 abyss kernel: scsi2 (12:0): rejecting I/O to offline > > > > device > > > > Aug 31 04:48:15 abyss kernel: raid5: Disk failure on sdm, disabling > > > > device. > > > > > > If you reread my message, I hope you will find a bit of clue: > > > > > > echo y > /sys/block/sdm/device/delete > > > (or something like that anyway.) > > > > > > That's needed for exactly this purpose: to remove a device which is > > > marked 'offline' in the kernel somewhere - to be able to re-add it > > > later - either with this script or just doing that magic > > > add-single-device > > > command manually. > > > > I should have listened sooner; I wasn't able to follow the paths in the > > re-add command, so I didn't do the delete. > > > > I did the delete (the above syntax is correct) -- ran the rescan script, > > it > > added it back - and the device is accessable again. > > > > Thank you Michael & Forrest. > > I just noticed that in /sys/block/sda/device/, along with the delete > file, there is a rescan file. I think that this is what I was thinking > of with the 2.6 kernel. Well, my worst fear was realized. My raid array barfed during rebuild. It errorenously kicked a few devices from the array. I/O was not rejected to any of the drives as was before; so I stopped the array. (mdadm --stop /dev/md0) -- and re-assembled it; mdadm -A /dev/md0 /dev/sda /dev/sdb /dev/sdc /dev/sdd /dev/sde /dev/sdf /dev/sdg /dev/sdh /dev/sdi /dev/sdj /dev/sdk /dev/sdl /dev/sdm /dev/sdn /dev/sdo /dev/sdp /dev/sdq /dev/sdr /dev/sds /dev/sdt /dev/sdu /dev/sdv /dev/sdw /dev/sdx /dev/sdy /dev/sdz /dev/sdaa /dev/sdab -f Just as before; mdadm: forcing event count in /dev/sda(0) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdb(1) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdc(2) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdd(3) from 337384 upto 2645518 mdadm: forcing event count in /dev/sde(4) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdf(5) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdg(6) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdh(7) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdi(8) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdj(9) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdk(10) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdl(11) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdn(13) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdo(14) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdp(15) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdq(16) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdr(17) from 337384 upto 2645518 mdadm: forcing event count in /dev/sds(18) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdt(19) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdu(20) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdv(21) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdw(22) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdx(23) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdy(24) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdz(25) from 337384 upto 2645518 mdadm: forcing event count in /dev/sdab(27) from 337384 upto 2645518 mdadm: /dev/md0 assembled from 26 drives - not enough to start the array. The 2 drives, one of which was sync'd -- and one that was rebuilding; won't join the array.