Jun 14 14:55:21 Tower kernel: BTRFS info (device sdi1): bdev /dev/sdj1 errs: wr 772370105, rd 641390571, flush 4473716, corrupt 4949264, gen 0
Jun 14 14:55:21 Tower kernel: BTRFS info (device sdi1): bdev /dev/sdi1 errs: wr 1603125, rd 1089630, flush 8890, corrupt 3731, gen 0
Both cache devices are showing a very large number or read and write errors, this means both have been dropping offline, see here for better pool monitoring.