HD crashed in Synology DS411+ (RAID5)

Question

Today I got notified by email as one HD (out of 4 identical 2TB HDs) in my Synology NAS was "disconnected". I don't have any physical access to the NAS at the moment, but tomorrow someone will replace the defect HD.

I have set up a weekly SMART fast test, which did not show any errors on last Wednesday.

My question is the following: In one notification email (regarding the worsened condition of system-volume) it says I should restart the NAS, because this will repair the system on startup.

• What does this exactly mean?
• Is it possible as the HD will work again?
• Would this resolve the RAID5?
• Do you suggest me to do this? Or should I just turn off the NAS until the HD is replaced?

Thanks in advance!

Answer 1

"Disconneted" is a vague term, you should probably consult synology documentation to find out what do they mean by this. It may actually mean that the drive was indeed disconnected (i.e. someone pulled it off by accident). In that case the drive itself should be ok, you can plug it back in and perform an array rebuild.

Of course, "disconnected" could mean more serious issues with your NAS: port failure on the backplane, partial controller failure, or some sort of catastrophic HDD failure, when the drive just suddenly stopped working. However, these issues are very rare. So, if I were to speculate, I would say that someone just pulled the drive.

If, however, "disconnected" means "failed", then it is not a good idea to rebuild RAID over it (especially RAID5). Drive failure detection techniques vary between vendors, and I do not know how Synology tells if the drive is faulty or not (it probably is based on failed reads/writes, bad block/reallocated sectors counters or something like this) but I would strongly recommend replacing it with the new one. This "failed" drive may still live a long and happy life in some PC, but not within RAID.

And finally, I do not think that rebooting NAS with failed soft-RAID (it is a soft-RAID isn't it?) is a good idea. I can't back it up with anything, but from my experience it just smells like trouble.

Answer 2

What does this exactly mean?

As @Tigran Balyuan indicated, 'DISCONNECTED' is a vague term to inform you the NAS couldn't communicate with one of the drives. It could mean the drive failed, the SATA (or power) adapters are not seated properly or the controller that's reading the drive is bad. No way to tell unless you physically inspect the device, but in most cases (in my experience with this message), the problem is due to the NAS and not the individual HDs.

Is it possible as the HD will work again?

Usually, if the HD is bad, the system will give you a more specific error; like the HD failed or has bad sectors. Restarting your system will give you a quick indication of whether or not the RAID integrity is in-tact. If it is and you get a more specific error about a particular HD, you can replace the HD and the system should automatically rebuild the RAID on next reboot.

Would this resolve the RAID5?

There's no indication that restarting or replacing a HD will resolve a problem. RAID5 just indicates your striping and mirroring configuration. RAID5 is generally resilient enough to stay functional up to 1 failed disk. If you need anything more fault tolerant, I'd suggest going to RAID1 or RAID1+0; note; going to RAID1 or RAID1+0 is generally more expensive, because you're sacrificing more disks for tolerance.

Do you suggest me to do this? Or should I just turn off the NAS until the HD is replaced?

Rebooting a NAS is basic a troubleshooting step. However, someone really needs to physically inspect the NAS. There might be light indicators telling you which drive is bad; or when the system is shut down, someone can inspect the connectors to verify solid contact. If the system sees a problem, you'll be in the same spot you are now. If you replace the bad HD, the system will automatically rebuild the RAID from parity bits.