Drives Randomly Missing, Can't Rebuild...

[youngnapoleon]

Hi guys! This is my first post & apologize a) for the length & b) for the likely large amount of mistakes I committed today. The main one being multiple reboots without saving syslogs – I’ve since read that this is probably my demise.

 

I got unRAID up & running last night & was happy with it, just threw on a few gigs of test stuff in a few different shares & created the primary users. Today was going to be spent doing The Big Transfer, buuuut I’ll sum the issue up in dot point:

 

(2x WD Red 4TB drives, one parity, one disk. 120GB SSD for cache & Sandisk Cruzer USB connected directly to the motherboard as recommended.)

 

• While moving files from an external HDD (using Unassigned Devices & Krusader), I paused the transfer with the Krusader GUI.

• Krusader froze, & then the unRAID GUI also became unresponsive. I waited over half an hour but could get no response from either.

• Fearing the HDDs may be encountering errors & leaving it longer may be damaging them (plus feeling hopeless after half an hour of nothing), I force restarted the PC.

• Back up, Disk 1 had the red cross next to it & was being emulated.

• I was fairly sure the data was still there. I removed Disk 1 & tried attaching it as an unassigned device & also to a Linux VM on my main machine – neither would read. This turned out to be due to my cheap ‘SATA to USB’ cable I’m pretty sure, but I’ve included this step just in case.

• With the parity seemingly fine & all files still accessible, I tried rebuilding by unassigning & reassigning Disk 1, stopping & restarting the array in between.

• This kicked off alright, but at ~0.6 percent, hit a read error on Disk 1 & aborted itself. I tried this a couple of times.

• I found a forum post saying to check the SATA cable – though it had failed before this, I HAD played with it, so made sure everything was snug.

• Same result.

• I tried rebooting a few times, beating my head against a wall. Weird things were happening. Both Disk 1 AND the parity drive would randomly appear as unassigned devices as if they were external. The mount option was there, but did nothing.

• After awhile, it got to the point where both Disk 1 & parity were simply not recognized. The flash & cache drives were both fine. Note that despite the parity acting up in different ways, I had never really done anything with it.

• Frustrated as all hell, I unplugged the SATA & power cables from the cache SSD & plugged them into the parity drive. Booting up, the parity is now apparently fine & dandy, AND Disk 1 is being recognized again, but still with the red cross. The cache is currently not connected at all.

 

I’ve gone from working happily to seemingly having lost the ability to read the 2 hard drives & everything in between today, so figured with the parity alright & Disk 1 at least being recognized, I should turn to the forums. I imagine all the fooling around I did was dumb before asking you guys, but I’ve never really had trouble I couldn’t fix myself until now. Every time I rebooted (except for that one fateful hard boot), I stopped the array & used the proper shutdown procedure.

 

Also note the machine is headless – I can SSH into it, so I guess it’s not a problem. I just really really don’t want to have to pull the GPU out of my main PC & stick it in this one again, so hopefully it’s nothing in the BIOS haha.

 

Anyway, sorry for the long post, thanks a lot for reading! Let me know if you have any suggestions

 

Edit: Couldn't help myself, rejiggered the cabling & attached the cache SSD. All drives showing fine except Disk 1. Would possibly the easiest way of fixing this/determining whether there's something wrong with the HDD itself just be to format it using Preclear or something?

Edited June 28, 2017 by youngnapoleon
Update

[trurl]

So far it doesn't sound like you have done anything really bad. Just don't do this:

54 minutes ago, youngnapoleon said:

Would possibly the easiest way of fixing this/determining whether there's something wrong with the HDD itself just be to format it using Preclear or something?

That doesn't really make any sense anyway because clearing and formatting are not even remotely similar, and neither would help.

 

Go to Tools - Diagnostics and post the complete diagnostics zip.

[youngnapoleon]

Thanks for your reply! I didn't read too much about preclearing, so I for some reason assumed it was similar to formatting 

After the successful boot with the cache I tried unassigning/reassigning once more, & despite it being ridiculously slow, it's been running for more than an hour now

Diagnostics attached, thanks again!
 

 

illmatic-diagnostics-20170628-2214.zip

[trurl]

I haven't looked at your diagnostics yet.

 

6 minutes ago, youngnapoleon said:

I tried unassigning/reassigning once more, & despite it being ridiculously slow, it's been running for more than an hour now

So is it rebuilding now? What do you mean by slow?

 

It will take a considerable amount of time to rebuild a 4TB disk, several hours, because it has to write every bit of the 4TB, regardless of how many files may have been saved. Parity doesn't know anything about files. It should show you the speed of the rebuild in Main - Array Operations and give you an estimated completion. Post a screenshot if you want.

[youngnapoleon]

Yeah, yesterday when I originally built it, it took about 8 hours. I can't remember the speed, but I think it was well over 100mb/s. Now it is going very steadily ~36mb/s. Having said that, over an hour in & no errors.

[trurl]

OK, I have looked at your diagnostics now and you have a bad disk:

Serial Number:    WD-WCC4E2ZZFYAH
  5 Reallocated_Sector_Ct   0x0033   133   133   140    Pre-fail  Always   FAILING_NOW 1963
  9 Power_On_Hours          0x0032   100   100   000    Old_age   Always       -       24

Looks pretty new too. You should replace it with another disk and return it.

 

17 minutes ago, youngnapoleon said:

I didn't read too much about preclearing

Now would be a good time to learn. Preclearing is the way most people test a disk before using it. Every bit of every disk must be good because every bit of every disk is required to rebuild a failed disk.

[youngnapoleon]

Damn, thanks for the heads up. Both disks were brand new. I've never RMA'd a disk before, guess it's time to start

 

Thanks again trurl!

[JonathanM]

Depending on HOW brand new and where purchased, I would pursue a return and rebuy rather than an RMA. RMA guarantees you will get a refurb disk, so if you can, return it instead.

[youngnapoleon]

17 hours ago, jonathanm said:

Depending on HOW brand new and where purchased, I would pursue a return and rebuy rather than an RMA. RMA guarantees you will get a refurb disk, so if you can, return it instead.

 

First time it's been used, but unfortunately bought them last September, not thinking it would take so long to get the rest of the hardware