Hello, I've got RH9 servers running in a few locations, with some 'valuable' data in a mirrored RAID partitions.
One such site recently had a disk crash in the mirror and we removed the disk from the system. However, a replacement was not readily available and the data was needed day-to-day. When we booted the machine it autodetected the single raiddisk, but the boot stopped because the second disk was 'moved' (i.e. the system had several other disks, which moved in system-ID because of the crashed disk's removal). All non-raid partitions could work with this just fine (because of the use of labels, thereby abstracting the disks from their system-id's [sda, sdb, etc]), but if the boot still hangs on the raid-detect, such 'fail-safe' measures become a bit academical. Is there another way (other than ensure there is only one raid-device and it's physical disks are the last in the system-id's; something which is unpractical when using three scsi-controllers and multiple disk-arrays) to ensure a system comes back up alright, even with a malfunctioning array? I guess hoping the raidtab can use disklabels anytime soon would be asking for too much... ;) Kind regards, - Joop Carels. You cannot teach anything to anyone, you can only help them discover it. -- redhat-list mailing list unsubscribe mailto:[EMAIL PROTECTED] https://www.redhat.com/mailman/listinfo/redhat-list
