Read errors on disk

[aspdend]

A little while ago I upgraded my setup to a new case with 24 hot swap bays so I could expand my array having reached maximum capacity in my existing case. At the same time, I added a second controller for the number of disks. All good. Everything transferred over smoothly. After making sure it was all behaving properly, I dug out one of my older disks that I had previously replaced with a higher capacity one, popped it in the new array and pre-cleared it. All went well, it passed. Did SMART tests etc everything passed. Today I got some read errors reported off the disk (67 iirc). I appreciate it was a risk to add in an older disk, but there was nothing wrong with it when it was replaced and the preclear/SMART testing all passed - in fact the SMART testing still reports it as passing. What would the best thing be to do with it now? I presume just clear it off, remove it and shrink the array back down. There is only a small amount of data on it (I made sure of that) and none of it is irreplaceable so I'm not worried about the data on it - just a bit loathe to bin a disk (definitely past it's warranty) unless I have to...

[JorgeB]

Please post your diagnostics: Tools -> Diagnostics

[aspdend]

3 hours ago, johnnie.black said:

Please post your diagnostics: Tools -> Diagnostics

Here you go

tower-diagnostics-20180815-1739.zip

[JorgeB]

Looks more like a connection problem, you could replace/swap cables/backplane and see problem doesn't reoccur, but if you want to remove there are two options, both described here:

 

https://lime-technology.com/wiki/Shrink_array#For_unRAID_v6.2_and_later

 

P.S. you should also update to latest release.

 

[aspdend]

I've been waiting to get my server relocated before updating - that was to do this weekend. The drive is on a backplane with 4 others - it#s essentially a Norco RPC-4224 http://www.norcotek.com/product/rpc-4224/ Which is why I wasn't thinking it's a cabling issue...

[JorgeB]

There are still cables, and although rare a backplane slot can also fail, only way to find out cor sure would be to swap slots with another disk, that would rule it out.

 

 

 

[aspdend]

Well I have spare cables so I could try that first then if nothing then I can swap drives. So what would I do after powering down to swap cables? Restart and run a parity check to see if any errors occur or just restart and see what happens?

[JorgeB]

1 minute ago, aspdend said:

Restart and run a parity check to see if any errors occur or just restart and see what happens?

Either way, but a parity check would likely be faster to detect any issues, just make sure it's non correct.

[aspdend]

2 minutes ago, johnnie.black said:

Either way, but a parity check would likely be faster to detect any issues, just make sure it's non correct.

Sounds good to me, as usual, many thanks for your sage advice, I will post an update when I know more

[aspdend]

On 8/15/2018 at 6:20 PM, johnnie.black said:

Either way, but a parity check would likely be faster to detect any issues, just make sure it's non correct.

Hi johnnie, I have now replaced the cable (after some struggles with the SFF 8087 connectors!) and no joy - still getting read errors. I then changed the port on the card and no joy - still getting read errors so I have changed the slot the drive is in - now the drive is on a different backplane using a different cable connected to the other card in the server and no joy - still getting the read errors.

 

I have attached the revised logs - any more advice? 

tower-diagnostics-20180819-1523.zip

[JorgeB]

No errors on latest logs, but if you've ruled out everything else you're left with the disk, remove or replace.