Hi - I have a RAID1, 2 disk mirror set that contains a drive that fails intermittently (or at least gets disk errors intermittently and so gets marked as failed). Obviously I want to swap the failing drive for a new spare (we arent running with a spare currently). So, I shutdown when the drive failed, replaced the drive and rebooted. What this got me was a degraded root partition (md0) and a non-working /export area (where the md1 device is mounted). md0 (root device) is hda1 and hdc1 md1 (/export) is hda3 and hdc3 hda1 and hdc1 are tagged as bootable, and of type raid-autodetect. On reboot the root partition gets auto-detected as raid, issues an error about "invalid raid superblock magic on hda1" and automagically sorts itself out mounting md0 in degraded mode. However, md1 does not. The only difference I can think of is that the md1 device is made up of two hd partitions that have not been set to the auto-detect type (the raidtab entries for md0 and md1 are identical bar the devices they use). When I reboot I get: Feb 10 18:58:05 bart kernel: md: invalid raid superblock magic on hda3 Feb 10 18:58:05 bart kernel: md: hda3 has invalid sb, not importing! Feb 10 18:58:05 bart kernel: md: could not import hda3! Feb 10 18:58:05 bart kernel: md: autostart hda3 failed! If I set the type to autodetect, will this device come up in degraded mode instead of failing like this? What I want is a system where, if a drive fails, someone can shut it down, replace the failed drive, reboot in degraded mode (at this point everyone else can start working again) and then partition the new drive and raidhotadd it back into the set. At the moment I've got a system thats got a more long winded process to repair (and requires knowledge of the RAID setup). Obviously having a spare in the set would be even better, but thats not an option at the moment. Thanks Marcus -- Marcus Williams - http://www.onq2.com Quintic Ltd, 39 Newnham Road, Cambridge, UK - To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
