I've been running unraid for a few months on a test rig, and I recently decided to move the data, then the bigger drive I had on my previous hardware NAS to the test rig.

 

The drive is a 4Tb WD Red, and it always worked flawlessly on the previous raid. I've decided to use it as parity drive (the other drives on the test rig are 1Tb drives), with the intention to slowly upgrade the rest of drives to all 4Tb Reds with time.

 

I don't turn on the test nas all the time, so I was quite surprised when I turned it on a couple days ago and noticed that the parity drive is disabled. I tried to parity check, remove it from raid, restart raid, stop raid and re-add the parity to rebuild it, but no avail. Have any idea of what I should try next? I'm including the logs.

 

griffonnas-diagnostics-20180401-0855.zip

griffonnas-diagnostics-20180331-2311.zip

No SMART for parity. Check connections SATA and power, and try again.

I tried changing the sata cable, just in case. The drive appears to be powered on, and the power cable was powering the previous parity drive, so I'm pretty sure it works. Should I do any special test with that drive, like taking it out and run some kind of diagnostics?

Post new diagnostics, drive was offline on the previous ones son no SMART report.

The drive figures as connected in the web interface... But still, as requested, new diags attached.

griffonnas-diagnostics-20180403-1116.zip

1 minute ago, griffon said:

The drive figures as connected in the web interface... But still, as requested, new diags attached.

Now it is, it wasn't on the previous diags, hence the need for new ones, SMART looks fine, if you already replaced the cable resync parity.

Alright, started the rebuilding process. Will update as soon it's finished. Thanks for now!

Running parity check returned 0 errors, but the disk is still marked as disabled. Should I "stop raid, remove parity disk, restart raid, stop it again, re-add, restart raid" now?

14 hours ago, griffon said:

Alright, started the rebuilding process. Will update as soon it's finished. Thanks for now!

 

17 minutes ago, griffon said:

Running parity check returned 0 errors, but the disk is still marked as disabled. Should I "stop raid, remove parity disk, restart raid, stop it again, re-add, restart raid" now?

 

So this is a little unclear. Did you rebuild parity, then run a parity check? Or are you confused about the difference between rebuilding parity and checking parity?

Quite so. I know parity check will only check if the data on drives is consistent with thte data on parity drive. As for rebuilding it, I thought that stopping the array, removing the parity drive from config, restarting it, then stopping, readding aand restarting would force a parity drive rebuilding process. Am I wrong? If so, please point me to the right process, pretty please. I don't want to lose the data on the drives...

Still not clear if you resynched parity or just did a read check, to resync parity, unassign parity disk, start the array, stop the array, re-assign parity, start the array to begin parity sync.

Just did the parity check. WIll do the rebuilding now. thanks for confirming the steps. Didn't really want to lose the data in there.

23 minutes ago, griffon said:

Just did the parity check.

Without a valid parity what you did was a read check.

Alas, something didn't go right. Rebuilding process seems to have stopped by itself, with errors. Attaching latest diagnostic...

griffonnas-diagnostics-20180404-1154.zip

Try swapping both cables/backplane slot and resync, if it fails again it could be the disk despite the healthy SMART.

I just tried to swap both power and data cables, and for good measure, I even plugged the parity drive on a secondary pci controller (it was already plugged in), then rebuilt the parity. Pretty much same results. I am attaching the diags after the deed, but at this point I'm  pretty sure something is wrong with the WD Red...

griffonnas-diagnostics-20180404-1401.zip

Well it does have some reallocated. I'm curious though why you were unaware of this.

 

If you look at the Dashboard, does it have any warning indicator for that disk?

 

Do you have Notifications setup?

 

 

30 minutes ago, griffon said:

I'm  pretty sure something is wrong with the WD Red...

Very likely you need a new disk, you can run an extended SMART test, it may or not confirm the problem.

44 minutes ago, trurl said:

Well it does have some reallocated. I'm curious though why you were unaware of this.

 

If you look at the Dashboard, does it have any warning indicator for that disk?

 

Do you have Notifications setup?

 

 

 

That's where the catch is. Last time I cold booted the NAS (before this chain of posts, of course - around 15 days ago), the dashboard considered the disk as "normal". I do have the notifications active. Now is marked as "faulty". The reallocated sector count wasn't yellow before the last parity rebuilding as well... So I guess the disk in a "pre-fail" state. Better replace it now, before it dies at the worst moment, I'd say.

 

The drive will only notice if some sectors have read issues if it performs a read attempt - so a parity check or an extended SMART test are good ways to find problems. Without full reads, a drive can have problems with sectors for months and years without anyone noticing them.

 

So any disk with data you care about should have an automatic process to regularly test the disk and report the result. Doesn't matter if it's the system drive of a laptop or if it's the data disks of your storage server(s). You want to make sure you get the warnings as early as possible to hopefully reduce the number of lost files.

 

If a single problematic sector shows up, it's very hard to know if it's a failing drive or just some outside situation affecting the drive when it was writing that sector - maybe an impact on the computer case or a power glitch. If multiple sectors shows up nearly in time, then I'd say it's more or less panic time to replace the drive.

 

The WD Red has waaaay too many reallocated sectors to just have had a tiny bit of bad luck with you knocking on the case or a power spike while writing. Something in the drive is marginal - and the drive could just as well have tens of thousands of broken sectors next week.

 

You could potentially use the drive as a third-tier backup target, but it isn't a disk you want as part of your array.

The Seagate drive whose serial number ends with BFHX (Disk 2) has even more reallocated sectors so when you start replacing 1 TB drives with bigger ones you ought to do that one first.