JorgeB

Yes, memtest is to try and confirm if it's a RAM problem, it won't fix the existing corrupt data, deleting the corrupt data should allow to remove the missing device from the pool, but possibly easier to juts backup, re-format and restore.

I had already moved your other thread here, so going to lock this one.

That suggests there's some issue with those drives, or a power problem.

Diags are just after rebooting, so not much to see, you can try this and then post that log after a crash.

Yes, copy super.dat to keep the disk assignments, or just re-assign them to the original slots and check "parity is already valid", dockers are easy to keep if you use the same paths, VMs might need some more changes, especially if passing-through hardware.

That is the SAS2LP known issue, so it should be fixed once you replace it.

Nothing obvious, I would try again after swapping the 16TB disks to the onboard SATA controller.

https://forums.unraid.net/topic/12114-lsi-controller-fw-updates-irit-modes/?do=findComment&comment=632252

In my experience that board is not very picky with RAM, I have some 1333 and 1600Mhz Kingston, as well as 1600Mhz Samsung RAM on my boards, it should work with most 1333/1600Mhz unbuffered ECC RAM, X10s are much more picky.

Diags are after rebooting, if it happens again grab them before.

You can use the NUT plugin, it allows master/slave(s) on the same UPS.

Please post the diagnostics: Tools -> Diagnostics

Error is logged as an actual disk problem, though since it passed the SMART test it's OK for now, keep monitoring that disk, you should also do something about the log spam.

There's a device missing from the pool, hence the current reported capacity: Data Metadata System Id Path RAID1 RAID1 RAID1 Unallocated -- --------- --------- --------- -------- ----------- 1 /dev/sds1 465.00GiB 2.00GiB 32.00MiB 464.48GiB 2 missing 61.00GiB 2.00GiB - -63.00GiB 3 /dev/sdr1 404.00GiB - 32.00MiB 527.48GiB -- --------- --------- --------- -------- ----------- Total 465.00GiB 2.00GiB 32.00MiB 928.96GiB This also shows some data corruption, and likely why the missing device failed to be removed: Feb 2 04:54:48 Jarvis kernel: BTRFS info (device sds1): bdev /dev/sds1 errs: wr 0, rd 0, flush 0, corrupt 7, gen 0 Feb 2 04:54:48 Jarvis kernel: BTRFS info (device sds1): bdev /dev/sdr1 errs: wr 0, rd 0, flush 0, corrupt 5, gen 0 Start by running memtest to check your RAM, then best to backup what you can from the pool and recreate, a scrub will indicate which files are corrupt.

If you have old disk1 intact, and since disk2 has no data, you don't need to rebuild, just do a new config with it, also filesystem status of one disk doesn't matter for the rebuild of another.

Emulated disk1 data is expected to be corrupt, do you have ?

Depends on the model, check endurance/TBW in the specs, also larger the capacity usually equals higher endurance.

SMART looks fine but full diags might give more clues.

Please post the diagnostics: Tools -> Diagnostics

Diags are just after a reboot, so not much to see.

Errors in multiple disks: Feb 1 10:24:47 XIO kernel: ata5: softreset failed (1st FIS failed) Feb 1 10:24:47 XIO kernel: ata5: limiting SATA link speed to 3.0 Gbps Feb 1 10:24:47 XIO kernel: ata5: hard resetting link Feb 1 10:24:47 XIO kernel: ata6: softreset failed (1st FIS failed) Feb 1 10:24:47 XIO kernel: ata6: limiting SATA link speed to 3.0 Gbps Feb 1 10:24:47 XIO kernel: ata6: hard resetting link Feb 1 10:24:52 XIO kernel: ata6: softreset failed (1st FIS failed) Feb 1 10:24:52 XIO kernel: ata6: reset failed, giving up Feb 1 10:24:52 XIO kernel: ata6.00: disabled Feb 1 10:24:52 XIO kernel: ata6: EH complete Feb 1 10:25:17 XIO kernel: ata1: softreset failed (1st FIS failed) Feb 1 10:25:17 XIO kernel: ata1: hard resetting link Feb 1 10:25:17 XIO kernel: ata3: softreset failed (1st FIS failed) Feb 1 10:25:17 XIO kernel: ata3: hard resetting link Feb 1 10:25:17 XIO kernel: ata4: softreset failed (1st FIS failed) Feb 1 10:25:17 XIO kernel: ata4: hard resetting link Could be the typical AMD controller issue, or a power problem, there are also some IOMMU errors related to the GPU, if it's the former upgrading to v6.9 might help.

Main difference is PCIe 3.0 vs 2.0, depends it your board can use it and it's worth it. Also depends on how many disks you have there, and how fast they are, but likely will make a bigger difference than the HBA.

Samsung SSDs and older AMD chipsets don't usually go well, try connecting it to the LSI, you'll lose trim support but still worth trying.

This is a device problem, and not that unexpected since it's way past it's predicted life: Percentage Used: 176%

It's the UD plugin, please use the existing support thread: