NAS disk / filesystem scrubbing on a schedule? Good idea?

[VirtualLarry]

After all the recent forum talk about Seagate drives failing, and knowing that my last little NAS backup was probably six months ago, I just did a full file-copy backup to my bigger NAS.

Well, the little NAS has two Seagate 1TB 7200RPM desktop drives in it, in a mirror configuration.

I don't think that I configured regular disk checks on that NAS. The drives are probably three years old by now, if not four. I ran into one file that wouldn't copy, for some reason. 1 of of 50,000. Not so bad. (It wasn't an important file, thankfully.)

But what if it had been important?

I'm honestly a bit surprised that the NAS wasn't able to piece together that file, given it should have had a copy on both drives. That means that I had the SAME bad sector on BOTH HDDs. Bad luck, I guess.

So, maybe enabling monthly or weekly disk scrubbing is a good idea after all?

Edit: It seems, that last time I upgraded the firmware, I DID enable SMART "Rapid test" weekly, as well as the "long test" monthly.

I'm also doing a disk surface scan.

SMART data for both drives shows no pending / re-allocated sectors. SMART shows "Good" for both drives, in green.

So now, this is even more a mystery. No SMART errors, no pending sectors, but yet, the NAS hangs and times out on one specific file, for some reason.

The admin logs did show a few warnings, to do a filesystem check, after the power went out. I guess I assumed that it would do that automagically, when powering back up, but it apparently does not.

I did do one of those, and it didn't show any errors or further information other than it completed successfully. So I guess I didn't have any filesystem errors.

 

[Insert_Nickname]

I'm honestly a bit surprised that the NAS wasn't able to piece together that file, given it should have had a copy on both drives. That means that I had the SAME bad sector on BOTH HDDs. Bad luck, I guess.

That's the problem with mirrored drives. If the file is defective for some reason (f.x. bad copy), and you're unlucky enough, its defective on BOTH drives.

 

[dave_the_nerd]

Yeah, RAID-1 isn't a parity check like you'd get with ZFS, it's just a straight mirror, good data or bad data.

IMO, a scrub or file system check should be a monthly maintenance task. Some people prefer more or less frequently depending on the size of their arrays and the amount of data they have.

That said, don't ask me when the last time I scrubbed my ZFS pools was... :heh:

 

[dave_the_nerd]

I'd also point out that if your NAS isn't backed up, you might as well not bother, since it's not like you can recover the bad data from somewhere else. Ignorance is bliss.

 

[VirtualLarry]

There was and is nothing wrong with the file, nor is there any bad sectors on either drive.

I'm more leaning towards this file triggering some sort of firmware bug or something.

I was able to copy it off the NAS using my Win7 laptop, and, eventually, the Win10 rig I was using for the backup, after about 10 minutes of retries.

Maybe a "weak sector" on the drive? But I don't think that's the case, either, because while the file was retrying, I was still able to access the drive, from another explorer windows.

Maybe it's not a NAS bug, but a Win10 bug?

 

[VirtualLarry]

There's no pending SMART re-allocated sectors on either drive, and both passed a surface scrub.

Edit: And I do replicate stuff between several NAS units.

 

[denis280]

this is my NAS.http://www.newegg.ca/Product/Produc..._re=nas_hard_drive-_-1Z4-0002-002R5-_-Product
Before it was a seagate and it did fail on me Larry,that's why i don't trust Seagate no more,to many fail year after year