JorgeB

There are a lot of checksum errors, that suggest a hardware issue causing data corruption, like bad RAM, or the NVMe devices are dropping alternatively.

You can if current one it's btrfs, but it would me much quicker to do and online replacement instead, you can replace a btrfs single device "pool", see here.

The errors on disk10 don't appear random, they are logged as media errors and the disk appears to be failing: ID# ATTRIBUTE_NAME FLAGS VALUE WORST THRESH FAIL RAW_VALUE 5 Reallocated_Sector_Ct PO--CK 100 100 010 - 104 87 Reported_Uncorrect -O--CK 092 092 000 - 8 197 Current_Pending_Sector -O--C- 100 100 000 - 16 198 Offline_Uncorrectable ----C- 100 100 000 - 16 Run an extended SMART test.

Looks more like a connection/power problem, you could swap that drive to another slot to rule that out, also a good idea to update the LSI firmware to latest (20.00.07.00)

Unfortunately nothing logged that points to the issue, could be a hardware problem, since it's a Ryzen CPU have you seen this?

👍

IIRC the SAS2LP doesn't perform very well with an expander, they are also not recommended for Unraid for a long time now, the expander works very well with an LSI HBA, you can see some performance numbers here.

Yes, update is very easy, similar to a board BIOS update.

Syslosg doesn't show the start of the problem but looks like one the cache devices dropped offline, run a correcting scrub and check all errors were corrected, more info here: https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=700582 If the scrub can't fix all errors or there are more issues post after reboot diags.

Yes, but it's a very old one, you should update to latest: 20.00.07.00

I remember seeing trim issues (and other connections issues) with some Samsung SSDs on some controllers, IIRC it's a general Linux issue, and if it's that it should be solved by using an Intel controller, they are also very picky with cable quality, but that doesn't appear to be the problem here.

Why mandatory? You just need to re-assign them, you won't lose anything.

Once a disk gets disabled it need to be rebuilt, since the emulated disk is now mounting and if contents look correct you can rebuild on top: https://wiki.unraid.net/Troubleshooting#Re-enable_the_drive

Not very common, you have the 9400-8i8e but it's usually pretty expensive.

Any LSI with a SAS2008/2308/3008/3408 chipset in IT mode, e.g., 9201-8i, 9211-8i, 9207-8i, 9300-8i, 9400-8i, etc and clones, like the Dell H200/H310 and IBM M1015, these latter ones need to be crossflashed.

Yes, if you transfer mostly large files, like media files.

Check filesystem on the emulated disk1, if contents looks correct after it's fixed, and only if that's true, you can rebuild on top, I would also suggest replacing/swapping cables on disk1 before rebuilding to rule them out.

It's possible to disable a drive on a new install, so you can rebuild or replace it, but you need to know the assignments, or at least with drive is parity (if there was only one) and which is the failed drive.

Tom already mentioned that next beta will have will have partitions on SSDs aligned on the 1MiB boundary, this will hopefully help with this issue, will require re-formatting them though.

if you're booting CSM try UEFI, or vice versa, sometimes one woks and the other not, no one knows exactly why.

That you can do, it uses a standard SFF-8087 connection.

Very easy, on Settings -> Network Settings -> Interface Rules You can boot with the GUI mode if no network is available before changing eth0, I would leave bonding disable for now.

Your NIC isn't yet supported, it should be on the next beta which is expected very soon.

Any relatively modern disk should be able read/write at more than gigabit line speed (for large files), except very close to the inner sectors for some slower models, you can use the diskspeed docker to have an idea of your disks performance.

Unraid requires a specif partition layout and more importantly a specific MBR signature, if you have or add parity, and since partition info is outside parity and recreated by Unraid, you should be able to rebuild one disk at at time (or two with dual parity), this is for example what most people who change from RAID to non RAID controllers also must do, and it always worked AFAIK, but you can check the emulated disk first and make sure it is mounting and data looks correct before rebuilding on top, other than that you'd need to backup each disk and have Unradi re-formatted them, then restore the data.