unraidhome-diagnostics-20260910-0653.zipHi, hopefully someone can help. I keep getting cache pool errors. So far I have tried different ssd's. I have tried different brands. I have tried moving different ssd drives to different slots on the mobo. I have completely deleted all data on the cache and rebuilt all of the data. I have checked all of my configurations and I can't find anything wrong. I ran fix common problems and was suggested to upload my diag file here and ask for help. When I installed my latest new ssd to use as cache I was able to go 3 days without pool errors, now I get them every 12 hours or so. I tried to reset the errors thinking it might be a one off issue but it still continues to happen. Can anyone point me in the right direction? Thanks in advance.

Edited September 14Sep 14 by mexicali

btrfs is detecting data corruption, start by runing memtest.

Ok, I'll report what I find

I've been running memtest all weekend. It's not reporting any errors. Now my log filesystem is at 100%

Post new diags please.

Thanks for taking a look. Here is the syslog and diag. Unfortunately they are after a reboot. I tried to access the server but it crashed and I had to reboot for access. Since my last post, I have checked that all drives are securely connected, removed and reseated all of the ram. I also stopped my Arr Stack just in case that may have been an issue. I still continue to get cache errors. I did see radarr come up in the diags, I uninstalled it and cleared any leftover files but I still get errors.

unraidhome-syslog-20260914-1956.zip unraidhome-diagnostics-20260914-1555.zip

Edited September 14Sep 14 by mexicali

The older syslog shows checksum errors on the cache NVMe and on /dev/loop2, which is the Docker image stored on that same cache drive. It also shows crashes in host programs such as sh, hdparm, and smartctl_type. That makes this look broader than Radarr or one bad SSD. Hardware or platform instability is still my main concern, but the logs do not identify which component.

Before changing anything else, please make sure you have usable backups of the cache data. The scrub found errors it could not repair. Also, was your weekend test boot-time Memtest86+ or the live Memory Tester plugin? How many complete passes ran, and were XMP or other memory overclock settings enabled?

I would next use BIOS defaults with XMP disabled and monitor the system for longer than its usual 12-hour failure interval. Set up remote syslog first, since the local log filled and the latest diagnostics were taken after the crash. If errors return, we can use that first-fault log to decide whether to isolate the RAM/board/power path or the cache drive’s connection path. I would not rebuild the Docker image or run btrfs check --repair until the underlying fault is controlled.

Thanks for the reply. Here is what I have tried prior to your reply. I deleted the docker img, deleted the cache, I stopped the array and removed the cache, started then stopped and reselected the original cache drive. Next I formatted the cache drive and changed it from btrf to xfs. So far I have not gotten any errors. It was a lot of probably unnecessary work, but I wanted to check as many boxes as possible. It is weird that this server gives me so many problems, I have a second Unraid server that has been up for more than 2 months with zero issues.

There's likely some hardware issue causing the problems; keep the syslog server running and post that if you have more issues.

Thanks, will do