Hi all,

 

A quick question, my server has been running just fine and I have it set to run a parity check on the 1st of each month, this normally runs fine but this time there is a problem.

 

Just after the check has begun, one of my drives throws up 123 read errors. It then continues the parity check which completes with 0 errors which does raise the question, How are there 0 errors found when there were 123 read errors?

 

I've run an extended smart test on the drive with passed with no errors also.

 

It is connected to a SATA card, a flashed Dell Perc h310 which has another 3 drives on it which are running without trouble and I powered off the server and reseated all the cables. Ran another parity check today which again showed up the 123 read errors.

 

I assume this means the drive is failing in some way but it isn't as clear as when they have failed in the past, the drive is still running, it isn't red balled or anything with there being no write errors (yet)....

 

Is there anything I have missed trying and if not when I slot a replacement in is it ok to just swap the drive and rebuild from parity like I would if it had died completely?

 

 

Thanks in advance

 

Tim

20 minutes ago, thebaldconvict said:

Is there anything I have missed trying

You missed posting a diagnostic after the read errors, but since you rebooted it's too late.

If its still showing the 123 errors after the additional parity check, then before rebooting, go to Tools - Diagnostics and post the complete zip.

Well it is currently showing the 123 read errors so I can do that

 

Once I knew they weren't going away I left alone and thought I would seek help from people who know this much better than me  

 

Anyway, please see attached.

tower-diagnostics-20170302-2006.zip

Hi all, 

 

I have a replacement drive arriving today (actually I got two so I can add a 2nd parity) so will set those off doing preclears ready to slot in but in the meantime do I need to do anything before swapping out this drive or can I literally swap them out and rebuild the array? 

SMART for disk5 looks OK, and the disk wasn't disabled, so I would just check the connections and try for another parity check. The read errors will clear after you reboot.

 

Bad connections are very much more common than bad disks.

Thanks Trurl but as above that is what I tried first of all and still got the read errors...

 

I have two new drives in there preclearing at the moment but will try another parity check before swapping it out,

 

I think if I am still getting the errors one of the new drives will replace this one and the other will become a second parity. I think the 2nd parity will take some of my anxiety away  

Just to hopefully finish this off, I hate threads that haven't got a resolution on them...

 

Thank you to Trurl, I had already checked the cables to all the drives in the server but when I was adding in the new drives I realised that disk 5 which was the one that was erroring was right at the bottom of the case and hiding away, I can only assume that I was careless and missed checking that one.

 

Anyway, unplugged and replugged a few times and turned it back on. Ran the preclears on the new drives and added one as Parity 2 which obviously loaded the disks up doing a rebuild, all good. I went one step further and then swapped a disk out for Parity 1 which again caused a rebuild and again 0 errors from drive 5.

 

All working well now I think and finally have twin parity.

 

Now my next challenge is use the spare 4tb to start moving data around to swap them all to XFS (and yes I have now realised I should have done it before adding the 2nd parity disk)