How to Efficiently Replace a Broken Disk from a Linux Software RAID 5

1. ddrescue the broken disk to a new, completely empty disc. At least zero the former raid superblock... (ddrescue -n faileddev newdev logfile; ddrescue faileddev newdev logfile)
If successful (if you have bad blocks that couldn't be rescued, you should preferably rebuild the raid.)
2. add the disc to the raid and remove the failed disc from the raid
3. go to the sysfs folder of the raid device, e.g. cd /sys/block/md0/md/.
4. put the new disc into the slot of the failed old disc:, e.g.  echo 2 > dev-sdc1/slot.
5. check the raid to be sure if all worked out well: echo check > /sys/block/md0/md/sync_action.

No comments:

Post a Comment

I appreciate comments. Feel free to write anything you wish. Selected comments and questions will be published.