I have a failing hard drive (/dev/sdb). It still works most of the time. I'd like to pull it from my system and replace it. Unfortunately, I do not know which of the hard drives on my machine is /dev/sdb/ I can rule out two of them because they have a different capacity. I have physically labeled all these drives with stickers on the outside. Once upon a time each of these drives also had a partition on them containing a file that identified the sticker that was on the drive. But these partitions no longer seem to have a a file system in them. Something happened over the last ten years or so.
Now I could make new partitions and put files in them with identifying information. And then I could unplug them one at a time, reboot and see which ones were still there. Then edit the identifying files to make them correspont to the outside labels. Trouble is, they contains partitions for a RAID-1. And when I disconnect one of them the RAID becomes defective, havint only one of its two drives. The computer of course will still work with one drive, but I'd like to avoid having one of the RAID partitions be effectively discarded or becoming desynchronised by this test. So is there a way of booting without trying to assemble a RAID? Or will its defectiveness be ignored it I don't try to read or write it? Is there a way of reassembling a RAID after it has temporarily been made defective? Would booting from an installation disk help? (assuming I can still use the machine's USB drives) If so, how do I stop it from messing with the existing paritions, since instalers seem to like wiping partitions clean? (Yes, the computer may need to be replaced. But then I'll have the same problem identifying my old hard drives when installed on the new machine.) Or is there some completely different way of accomplishing what I want? For example, removing the disk drive and connecting it to my laptop (a different computer) using a SATA-USB interface? Would it try to assemble the RAID if I just plug it in that way and have the same problems? -- hendrik _______________________________________________ Dng mailing list [email protected] https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
