This is now the third time I have to backup, reformat, and restore my cache pool due to BTFRS filesystem issues. Logs attached.

 

Any idea what's causing this? I thought it was a bad cables as I would have ECC error count on the SSDs move from 0 to 1 and back regularly,  so I replaced my SATA cables. Then one in particular still kept throwing that error a lot so I bought a new SSD to replace it in the pool, but it's now my understanding that's just some kind of bug with MX500 SSDs and I'm still having this issue.

 

I really like the idea of having SSD drive fault tolerance with a pool, so I really don't want to go to a single drive XFS. Plus my OCD kicks in when I see that exclamation point on the shares tab telling me my cache shares aren't protected.

 

 

syslog.2

diagnostics are always better than a syslog

1 hour ago, Squid said:

diagnostics are always better than a syslog

 

woops, ok. here it is.

unraid-diagnostics-20220716-1447.zip

Repeated BTRFS issues can be caused by memory that's a little flaky or other marginal issues that don't effect other operations, it seems to be more sensitive to hardware that isn't 100%. How long has it been since you ran 24 hours of memtest86.com USB with no errors?

this is the first i'm hearing of memtest86. my motherboard doesn't support ECC ram, so i'm limited to non-ECC. could that be a factor here?

 

to me it just seems the raid1 cache drives were not balanced and just threw a ton of errors without a way for me to rebalance. i wound up swapping the older SSD with the other older SSD and let unraid rebalance, and i've been fine ever since so i guess this wasnt a filesystem corruption. i'm still not 100% sure what happened.

If you don't have ECC, and you don't boot UEFI, then the builtin memtest on the Unraid boot menu will work.

 

memtest86 was mentioned because it works whether you meet those conditions or not.

 

You should do memtest.