[reopened] Tricky disk problem (unRaid 4.7)

[pantheis]

I noticed my unraid box responding slowly to some file access and checked the system log. There were disk errors showing on the original data disk I had purchased. Checking the SMART report shows about 1400 sectors pending relocation, but no actual relocation events yet. Trying a number of offline utilities to get the disk to recognize if it needs to spare them out or not does not necessarily seem to have worked, so it seems it's time to replace the disk, or at least get it offline into another machine to work on as I need to get my unraid array back up.

 

Please note, even with the read errors in the log, there have been zero write errors noticed by unRaid and the array did not offline itself, nor did the disk red ball. I run automated monthly parity checks and those have all shown zero errors.

 

I currently have 4 drives attached, all 2TB drives. When purchasing the replacement drive, what I'd like to do is upgrade the parity drive to a larger size (4TB likely), then use the 2TB that was the parity drive to replace the failing disk. I'd hate to spend the money on just buying a replacement 2TB drive when it's just a tiny bit more to open up larger size disks in the future. I also realize that this will require upgrading to the 5.x unRaid to support the larger disks.

 

What is the best way to make the above changes, assuming that the parity drive is correct, but one of the data drives is suspect?

 

Thanks!

 

I can provide logs if needed, but I'm pretty sure the disk is failing.

 

--edit--

 

I had removed the failing disk from the array, started and stopped the array, then readded the disk to force it to rebuild the data on the disk from parity. The rebuild finished without any errors showing up, but after a bit I started to see errors again. I've attached my syslog and the front page of the web UI as requested just to get a second opinion on it. I've already ordered a replacement drive as the drives internal long self-test indicated possible failed sectors, but if there's something else going on as well I'd like to know about it.

 

Thanks!

syslog-2015-05-16.zip

20150516_smartreport.txt

[trurl]

There is a procedure called swap disabled just for this purpose. New parity is put in, old parity is put in place of failed drive, and unRAID copies the old parity to the new then rebuilds the data on the old parity.

 

It has been discussed a few times in recent memory on the forum. See my sig for search tips.

 

Having said that, you are on a version that most people on the forum will not be currently using. I have never done the procedure, and I don't remember much about 4.7.

 

After doing your research, I would suggest you get further guidance from some of the more experienced users here.

 

Just to save everyone some time, and so we can make sure we fully understand your situation, please go ahead and post both the syslog and smart report for the failing drive, and also a screenshot of the main page.

[pantheis]

Thanks for the information. I'll look up the procedure called swap disabled. Once the current offline utility finishes (probably overnight), I'll get the SMART and syslog reports posted as well as a screenshot of the main page.

[trurl]

What exactly are these "offline utilities" and how are you using them?

[pantheis]

I'm using Spinrite to run a read pass across the suspect drive.

 

--edit--

 

I've ordered a 2TB drive to replace the suspect drive for now. While I'd love to take this opportunity to upgrade to a larger parity drive for future expansion the drives are a little more expensive than I'd like to spend right now. This will also minimize the complexity of repairing my array right now by not having to also deal with upgrading to the next version of unRaid.

 

Thanks for the information on the swap disabled option. When I'm ready to actually update my parity drive, I'll now know how to go about doing it.

 

[pantheis]

Trurl,

 

I've added the syslog and a screenshot of the front page of the webUI as requested. I'm still pretty sure it's the disk itself, but I wanted to get a second opinion.

[trurl]

Even though you are not on that version, I think you can get a smart report for posting by following the v5 help in my sig

[pantheis]

Ah yes, sorry about that, forgot to include it.

 

Added to the opening post.

[trurl]

Some wiki search results for you

[pantheis]

Thanks, I did find that information before, but won't need to use it this time. I'm not upgrading to a larger disk at this time due to the complexities that will be involved in dealing with a disk that's potentially failing and having to upgrade unRaid versions all at once.

 

My plan right now is to spare out the drive with the replacement, then hit the old drive with a couple passes of preclear to see what happens.