Array disk & cache errors

[localhost]

Hi UNRAID community,

 

I've been having some problems for the last week or so that I've been unable to resolve. To try and avoid a wall of text I'll add some bullet points for things as they happened:

 

• Parity 2 went offline showing 2000+ errors. I could not spin the disk up and a reboot would make it disappear entirely. I assume disk failure.
• Swap new disk in, rebuild sucessfully - all seems resolved
• Test 'failed' disk - passes all extended tests
• Transmission stops being able to write to cache (BTRFS - SSD), errors go from I/O error to read-only error
• I assume a FS corruption, decide to switch cache to xfs. Reformat & works for a bit.
• Parity 2 goes offline - exactly as before
• Swap Parity 2 disk again to test the new disk - after reboot cache has no filesystem and needs formatting

 

Now when I try to rebuild the array it always fails after a few minutes on parity 2, and the cache keeps going unreadable. I haven't pulled the cache drive for tests yet but it passes SMART. I have also run a couple of passes of memtest just incase, which passed. Any help would be much appreciated.

 

Edited August 29, 2019 by localhost

[JorgeB]

You have what appears to be connections issues on both parity2 and cache devices, recommend replacing both cables on them, then run xfs_repair on the cache device.

 

 

[localhost]

Dayum, we were on the same page just then. After making the post I went out to my shed and grabbed some sata cables as a last resort. Have swapped the cable for parity 2 and its now rebuilding.

 

I never understood this, even almost a decade ago when I just to work in IT, how the hell does a sata cable just fail. Its obviously not the first time I've seen it but I'm always just left in disbelief. Its a low power signal cable that hasn't been touched, and just fails...

 

Anyway i'll see how it goes and if I have anymore problems with the cache i'll do the same. I refuse to believe 2 cables have failed simultaneously...

 

Thanks for your time

[itimpi]

It might be worth using something like deOXit on the connectors to improve conductivity.   This is one area where one can get degredation over time.

Edited August 29, 2019 by itimpi

[localhost]

Thanks for the suggestion, its a brand new disk and cable now so I should be good in the short term but will consider that for sure.

[JorgeB]

5 minutes ago, localhost said:

After making the post I went out to my shed and grabbed some sata cables as a last resort. Have swapped the cable for parity 2 and its now rebuilding.

That's good, but I meant replace both cables on each device, sometimes power cable can cause similar issues, but for now see how it goes with just the SATA cables.

[localhost]

Its all rebuilt and back to normal now, I will keep an eye on it. Thanks