I've never seen this happen.

 

I have a friend's server here (running 6.1.9); he had a data drive redball. When I got it to my place, all drives appeared green after I re-arranged the sata ports they were plugged into (was making room for another replacement drive).

 

When the array was started, it automatically started parity. Now I'm seeing, 6.2% in that it's status is complete. It's reporting 768 errors on the parity drive and a larger number of errors during the initial part of the parity check. Now all I see is a red X beside the parity device and "Parity device is disabled" when I hover over it.

 

What should I do?

 

Attached are the diagnostics.

aqua-diagnostics-20161030-1954.zip

Are you sure you didn't leave something out in your description? It should not be possible to enable a disabled data disk without rebuilding it or setting a new config.

I glossed over the fact that I swapped motherboards (to the exact same model) and power supply. I added a controller card too. When I booted the system for the first time since I got it home, disk 3 that had been redballed was appearing green again, not having touched anything within unraid itself.

 

It still appears that I can access data, but I'm not sure how to proceed. I thought disk 3 was the trouble drive, now unraid disabled the parity drive on me.

This looks more like a hardware problem.  At least, that is where I would start from. 

 

Double check all of the power connectors and SATA connectors. 

 

Run memtst for twenty hour hours.  (You have a new MB and PS (and memory?).  It would not be a bad idea to burn them in a bit...)

 

There is also the possibility of a bad PS.  (Does the new supply have a single of dual 12V rail?  You can tell by looking at the spec's.)

Hi Frank,

 

I'll start memtest now. It was on my list anyway. I was troubleshooting some systems here and just swapped out the motherboard & psu (not because of anything with this system but to help narrow down an issue with another). The motherboards are identical; old, used Asus M2A-VM boards and the power supply was swapped from a within-the-last-year new PSU to a more solidly built (feeling) 80+ bronze new EVGA single rail PSU. Memory stayed the same. I did update the BIOS.

 

I mentioned adding a controller; a SIL 3512 for the sole purpose of having another port to pre-clear the drive that was intended to replace the redballed disk 3.

 

I had moved one of the data drives to the controller, freeing up a motherboard port for the drive to be pre-cleared.

 

The issue now is knowing which drive to trust. Presumably disk 3 didn't redball for no reason. (It had been that way for some time). Now unraid is taking a fit with the parity drive. I'm not sure which to trust - but I feel as though I'm teetering on the edge of data loss here. I don't want to take a wrong step.

 

I'll report back after 24 hours of memtest.

 

After the memtest, would it be advisable to try and run another parity check (it may try to force me) or should I focus on replacing disk 3? Or the parity drive it's now taken issue with?

 

 

 

 

Did you blow out the SATA connectors to be sure there wasn't any dust and dirt in them? 

 

I personally would be using that SIL controller for the preclear operation rather than hanging a data disk on it.  Some of those cards (that chip set has been used by a lot of Chinese houses to make controller cards) have given folks issues in the past as unRAID operations can really stress the electronics. 

 

I've been away for a bit, but I did leave memtest run for more than 24 hours and no errors were found. I'll start it up later today after re-checking all the sata connectors and see what happens.

Well, I guess I had been mentally avoiding this project until I got a call about the status of this server. Somehow a month has slid by. As per the advice given, I blew out all the SATA cables, returned the data disk to the motherboard port and actually removed the SIL controller / drive to pre-clear for the sake of troubleshooting.

 

I started up Unraid and am still met with a red X beside the Parity drive.

 

I don't get it. One day Disk 3 is red balled, then it's not, then Parity begins now Disk 3 is green and Parity has the X. With two drives in question I feel I'm in a rather precarious situation.

 

If it were you, how would you proceed?

 

Edit:

The parity drive is showing 0 for reallocated, pending sector, unforrectable, CRC error count etc. I just ran the SMART short self-test and it completed without error. On the surface, the drive looks fine, yet it was disabled. How do I get Parity back without jeopardizing the questionable state of Disk 3?

Get a new diagnostics file and attach it to your next post.

I had started following the sequence in this thread: https://lime-technology.com/forum/index.php?topic=46113.0 until I read ahead and saw that it was (obviously, hindsight) going to start a parity check again.

 

I left parity unassigned for now, started the array and just so happen to have enough room left on my server to copy off all the data so I can work on it without fear or losing any.

 

Attached is the latest diagnostics.

aqua-diagnostics-20161203-0836.zip

All of the data is now backed up. I feel better about tinkering with it now. Again, everything is back to being plugged into the motherboard ports, no SIL3512 card in there.

 

I rebooted the server, re-assigned the parity disk which of course triggered a parity rebuild upon starting the array. Parity is now being reported as valid. It's telling me a parity check hasn't been done so I have that running now. 

 

My plan after the parity check completes is to run an extended smart test on each drive. I see that disk4's smart status was flagged: Reported Uncorrect - 20. If everything checks out well enough, I'm going to put another sata card back in (probably not the one I used last time) and hook up the other 1TB I have here to preclear it.

 

At the same time, I'd like to preclear a 640gb SATA drive I have here to replace a 400GB IDE drive currently in the system.

 

I guess the question is - is there anything else I should be checking? I'll be pulling out the questionable disk3 that redballed on me, but beyond that I'm not sure what else I can do to ensure trouble-free operation of this server.

Attribute #187 is not one that every drive manufacturer reports and I didn't see listed when the problem(s) occurred.  Try googling "SMART 187 Reported_Uncorrect"  and look at some of the results.  But that drive has been in service for over 6 years...

 

No server will ever be 'trouble-free'.  You have to always be keeping an eye on it.  Be sure that you set up the Notifications and look at the e-mails that you get.  That way, you stand a very good chance of only having minor issues rather than having those minor issues that have been ignored that result in a major problem.  You could move on to version 6.2.4 and  setup dual parity which would mean that you would have to have three disks fail before you would lose data.  But that extra protection is worthless if you don't monitor the server a regular basis!

Thanks, I'll look into it.

 

Yeah the drives in this rig are no spring chickens. They lived their lives in a raid 5 array prior to this.

 

So the parity check completed and apparently there are no errors, so not sure what the deal was with that. In any case, on the surface the server seems to be happy. I still don't trust disk3 as it's actually redballed twice now. I should proceed with pre-clearing the (old) replacement and see where that gets me.