Been running this server for close to 5 years problem free only a few reboots a year, so as with anything working a little rusty on setup and how things work.

While I was at work today got reports from family plex was down, got home and found all my dockers not loaded and seems the cache has gone read only. Cache is a two SSD BTRFS RAID.

Running short and extended SMART tests on both cache drives passing with no errors. Lots of space on cache available and Vdisk is only 51% used.

Not an expert at scanning diagnostics but looks like around 7:44 this morning the BTRFS had a error and switched to read only

<Aug 9 07:44:01 REMOTE kernel: BTRFS error (device sdb1): block=3999596822528 write time tree block corruption detected

Aug 9 07:44:01 REMOTE kernel: BTRFS: error (device sdb1) in btrfs_commit_transaction:2537: errno=-5 IO failure (Error while writing out transaction)

Aug 9 07:44:01 REMOTE kernel: BTRFS info (device sdb1 state E): forced readonly

Aug 9 07:44:01 REMOTE kernel: BTRFS warning (device sdb1 state E): Skipping commit of aborted transaction./>

Would really appreciate if anyone could point me in the right direction on this or point out any issues they see in attached diagnostics file I just exported.

Lucky it is my Friday today so can try to get this sorted. Setup this server close to 5 years ago using Spaceinvader One video's and besides adding a few HDD's and cleaning dust it's just worked. Have to refresh memory and get back up to speed on Unraid configuration :D

remote-diagnostics-20250810-0012.zip

8 hours ago, BackupAddict said:

write time tree block corruption detected

This typically means bad RAM, start by running memtest.

Thank-You for the advice. Was on my list of possibles but helps to not be randomly trying stuff if more experienced folk can point one the right direction.

Server was a little grumpy shutting down as it hung trying to unmount my cache, had to get monitor direct on server and powerdown from cli. no biggie needed to drag monitor in there anyway for the memtest.

Running now will see what it finds, if no errors on first pass how many passes do people think rules out mem problems?

edit. nvm, within 40 mins of memtest got a fail error, now to isolate bios settings, what stick or mobo/slot or if need to re-seat. Some backup RAM that should be good, but dropping from 32GB to 16GB hurts lol

Thank-you, assuming I don't have two things true at once likely resolved :)

Edited August 10, 2025Aug 10 by BackupAddict