When a drive fails preclear...

[TyantA]

We do the test to let us know if drives are about to fail. 

 

I'm working on a backup unraid box using older drives from my main server. I've run 7 through pre-clear and all of them passed except for 1. Now what? 

 

Skimming the Smart data, there's one row highlighted yellow: current pending sector with a value of 1. 

 

Non highlighted things that stand out are raw read error rate of 3, yet everything else is 0 (that looks like it would be of concern if it weren't). To use the drive or not - that is the question. I was planning on keeping one drive as a hot-spare anyway. Maybe that's this drive? Or maybe it's better to put it in the array and have one that did pass the test ready to go in? 

 

How do you decide when it's time to retire an old drive? 

[trurl]

For a small number of pending, you can try another preclear to see if it will reallocate the pending.

[TyantA]

It was the post-read that failed. I went to start it again and it asked me to resume. Figured that wasn't the point so a fresh preclear has begun! Will report back. Thanks. 

[trurl]

More than one person (including me) have discovered bad RAM when post read failed. Maybe you should cancel the preclear and eliminate that possibility before doing anything else since bad RAM is a major problem for the whole system. Then you can try the preclear again to see if the disk will pass.

[TyantA]

Interesting. Both that and the fact that pre-read already failed. 

 

Edit: like, 55 seconds in. 

 

I shut the array down, pulled the sata & power cables and noticed the WD drive in question's Sata cable had little to no resistance. It's possible the cable was quite loose. 

 

Still, I unplugged all the drives (no point in running them) and fired up memtest for an overnight run. We'll see what that comes back with, then I'll attempt another preclear with a more secure SATA cable setup. 

Edited December 11, 2018 by TyantA

[TyantA]

Ok so... I ran memtest for many passes and it checked out. 

 

Now when I fire up Unraid, I no longer see the preclear option under tools. Also, the drive shows up as precleared under unassigned devices so I can't start another preclear from there either? 

 

I tried reinstalling the preclear plugin but it's still not showing up. How do I run another cycle on it? 

 

Edit: I found the plugin error tab which noted an error for the preclear plugin so I deleted it and re-installed. That got it back in the Tools menu so it's running ... sigh. and it just failed. It's saying the smart overall health self-assessment passed but I do see current pending sector: 20. 

 

It's weird. It went from getting to the post-read stage the first time and now it can't even last a minute of pre-read. Like, literally a second after it started. 

Edited December 12, 2018 by TyantA

[Vr2Io]

On 12/11/2018 at 9:59 AM, TyantA said:

How do you decide when it's time to retire an old drive?

If drive fail and fail again, I will retire it.

[trurl]

11 hours ago, TyantA said:

current pending sector: 20

I would retire it since pending is increasing instead of decreasing. A pending sector is one that needs to be reallocated. The drive should be able to reallocate the sector and then the reallocated would increase and pending would decrease. A pending sector can't be reliably read and so could compromise a rebuild since all bits of all disks must be reliably read to reliably rebuild all bits of a missing disk.

[TyantA]

Fair enough. Will do. I'll relegate it to some single-drive task in another system until it completely fails.