I recommend sfdisk -d to rebuild the partition layout - see why at <http://mark.foster.cc/kb/raid-rebuild.html> -mdf
> From: Samuel Flory <[EMAIL PROTECTED]> > To: [EMAIL PROTECTED] > Subject: Re: Recover Raid 5 of Red Hat 7.3 > Reply-To: [EMAIL PROTECTED] > > Rus Foster wrote: > > >On Thu, 13 Mar 2003, Hong Tian wrote: > > > > > > > >>If one disk of Raid 5 is found bad, Could I just replace the bad disk and > >>recover the system and data? Or Should I re-install the Linux system from > >>scratch again and recover the data from the backup? > >> > >> > > > >The idea of RAID-5 is to survive this. Assuming you've only lost one disk > >then you should be able to plug in the spare and they system should > >rebuilt the data from the paratiy bits on the other disks. > > > > > > Well that might occur on a scsi raid controller with the drives on a > saf-te back plane. With software raid you will need to replace the > failed drive. Create new md partitions on the drive, and do a > raidhotadd to the md partitions. > > Example I've got disk sda-sde, all disks are the same size and have the > same partitioning scheme, disk sdd has failed, and been replaced. I > have 3 md partitions md0/sd*1, md1/sd*2, and md2/sd*3 > > -fdisk -l /dev/sda (figure how the disk should be partitioned) > -fdisk /dev/sdd (configure the new drive as the other drive) > -cat /proc/mdstat (Check out what drive goes where.) > -raidhotadd /dev/md0 /dev/sdd1 > -raidhotadd /dev/md1 /dev/sdd2 > -raidhotadd /dev/md2 /dev/sdd3 -- redhat-list mailing list unsubscribe mailto:[EMAIL PROTECTED] https://listman.redhat.com/mailman/listinfo/redhat-list
