I have 2 Samsung 860 500GB SSDs set up as mirrored cache pool.This has been running fine for about 9 months, however since a few weeks ago I got this error:

Unraid BTRFS error (device sde1): block=84020346880 write time tree block corruption detected

At the same time, my cache pool would just not be usable, and the logs flood.After a restart of my system, everything is fine again.

a scrub also shows no errors, btrfs check --readonly also shows no errors

I have also tried reformatting the pool, but that also didn't seem to help.

What can I still check/do to get rid of these errors?I will replace the SATA cables this afternoon, but wouldn't that also show itself in different manners?Like high error counts on anything?

The error count in the unraid main screen is still showing 0 errors when this happens.