Roberto Pereyra spake the following on 4/26/2007 5:08 AM: > 2007/4/26, Akemi Yagi <amyagi at gmail.com>: >> On 4/25/07, alex at avantel.ca >> <alex at avantel.ca> wrote: >> > On Wed, April 25, 2007 11:54 am, Roberto Pereyra wrote: >> >> > hdc3 appears to be the problem. Try using mdadm (man mdadm) query >> mode to >> > find out more info about md1 and /dev/hdc3. Are you sure /dev/hdc3 >> is set >> > to be linux raid autodetect (use fdisk to find out). You can also try >> > adding hdc3 to the array with mdadm - see the man page for details. >> >> I recently had a similar problem. One disk was reported to have >> failed. I ran mdadm to add it back. The reconstruction took a couple >> of hours (500GB) but it seems to be working now. So, as the above >> advice suggests, read man page for mdadm and try to recover the lost >> partition. >> >> Akemi >> _______________________________________________ >> CentOS mailing list >> CentOS at centos.org >> http://lists.centos.org/mailman/listinfo/centos >> > > > > Hi ! > > Can I force the reconstruction without risk with this command ? > > mdadm --assemble --force > If you have to force it, I wouldn't do it. Why not format the partition as ext2 with mke2fs -c to test it thoroughly ( or even -cc for a deeper test). Then you can redo it as Linux raid and re-add it into the array -- MailScanner is like deodorant... You hope everybody uses it, and you notice quickly if they don't!!!!