Synology NAS Manual Raid Repair
This is a technical note for a solution to a re-occurring problem with my Synology DS1815+ NAS and DX513 expansion unit.
PROBLEM: IO Errors appear on one or more hard disks. Array enters degraded mode. Re-seating the drive puts the Synology web interface into an known state.
Synology logs the message "An I/O error occurred with the hard drive...." Testing the hard drive, however, shows that it is functioning normally. Removing, dusting and re-seating the drive appears to cure the I/O error problem, as the drive appears "normal" in the Storage Manager interface.
Re-seating the drive puts Synology's "Storage Manager" interface into an unknown state: The volume appears degraded but all disks appear as "Normal". Screenshot below. There is no option to re-sync the volume or restore a drive to an array. Contacting Synology's technical support group proved unworthwhile.
SOLUTION: it is necessary to manually re-attach the missing drive to the array. This must be done via console using mdadm.
Step 1: locate the degraded array using cat /proc/mdstat. In this case it is 'md3'.
Step 2: Figure out which drive is missing. In this case it is 'sdf6'
Step 3: Manually add the missing drive to array using:
mdadm --manage /dev/md3 --add /dev/sdf6
Step 4: Confirm raid repair triggered using Storage Manager interface.