This documents a typical sequence of multiple drives failing in a RAID5, this time in Iomega ix4-200d.
I’ve received an automatic email from the dashboard saying
Data protection is being reconstructed. Data is available during this operation, however performance may be degraded.
After that, the NAS started ‘Data recovery procedure’ [and then came another message]
Drive number 1 encountered a recoverable error.
[and] NAS started recovery procedure from the scratch. Even though that mentioned drive has failed, everything worked fine untill yesterday [when] new message … said
Storage failed and some data loss may have occurred. Multiple drives may have either failed or been removed from your storage system. Visit the Dashboard on the management interface for details.
Is there any way to recover at least some data if it is NAS that failed?
Depending on the condition of the disks, Home NAS Recovery may or may not be able to extract data from it.
Maybe there was a spare involved, as data protected is being reconstructed in a RAID5 can only refer to a rebuild of the array. The rebuild happens either when a defective disk was replaced or when a hot spare kicks in after one of the active array disks fails. There is another variation to that tune, not really obvious, which is a transient failure causing one of the disks to drop out of the array momentarily, then report back online and be accepted back in the array.
Anyhow, while rebuild is in progress, it turns out the second drive in the array is unreadable. This halts the rebuild. The error is at first deemed recoverable, and the rebuild is retried. However, the error recovery is not successful and the second disk (#1) drops offline. With two drives offline, the data is no longer accessible.
For our data recovery software, the logical reconstruction is not a problem given that the disks are still readable enough. This may be a problem though. In worst case, the disks need to be cloned to a blank new disks, and clones then used for recovery. The NAS will not accept the clones because it already has recorded the disks as “failed”, and cloning the entire disk content also clones the “failed” marks for respective disks.