JorgeB

That is not the mover, if the share is set to cache=yes and it runs out of space data will overflow to the array: https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=537383

Mover runs on the set schedule: Settings -> Scheduler

Writes are always done to one data device plus parity, and they will be limited by the slowest of the two. Tthough there's no official array SSD support yet, they work, and work faster that disks, if you use good SSDs, I get close to 400MB/s on my small SSD array.

Don't understand what you mean, if you mean the emulated disk7 it's still part of the array, if you mean the old disk it will have whatever data it was there.

Basically, no.

Some data loss is expected: Best bet to recover as much data as possible is to clone disk1 to a new disk using ddrescue, then do a new config with the clone and the remaining disks, run an extended SMART test on parity, if it fails replace it also.

Not likely, look for a BIOS update, newer/different kernel might also help, you could try v6.9, post the output of cat /proc/interrupts to see what's using IRQ16, but based on the trace it's being used by a USB controller and one or both NICs, so you could have issues with those if in use.

I would think it would call shutdown directly, that script if just for handling ACPI events, like a button press.

NIC is bound to vfio-pci so it can't be used by Unraid, delete /config/vfio-pci.cfg from the flash drive and reboot.

You could always upgrade to a dog One solution is to rename the script that handles ACPI events, e.g.: mv /etc/acpi/acpi_handler.sh /etc/acpi/acpi_old.sh This would need to be done after every boot, with the go file or with a user script, also note that a long button press will still do a hard shutdown, as that's not software controlled. P.S. pretty sure that ACPI script won't be missed by anything else but not 100% sure.

There's something reading from disk2 in the diags posted.

You can enable turbo write for better performance, at the expense of all disks spinning up for writes.

This, and you can change them to the HBA, if it's really a true HBA in IT mode, no personal experience with that model but I believe it's been mentioned before by other users.

It is normal.

You can try this and then post that log after a crash.

This is normal, you can ignore as long as you checked "parity is already valid" Then you'd need another disk.

It won't hurt to try. -Tools -> New Config -> Retain current configuration: All -> Apply -Check all assignments and assign any missing disk(s) if needed, including the new cloned disk as disk5, and for now assign old disk5 as disk4 (you just need it as a temp disk), remaining disks should all be in the original positions. -IMPORTANT - Check both "parity is already valid" and "maintenance mode" and start the array -Stop array -Unassign disk4. -Start array (in normal mode now), ideally disk 5 and the emulated disk 4 would mount now, but that would e a little miracle, if they don't run a filesystem check on both -whatever is on emulated disk 4 is what's going to be there after a rebuild, so if it mounts (independent of disk 5 mounting or not) and there's some data there you can proceed with the parity swap.

BTW, total usable space will be 2.25TB

btrfs raid1 with an odd number of devices gives incorrect free space, and it will change as you fill it up, that's a known btrfs bug, total size e used space should be correct, strange about the pool size.

https://www.amazon.co.uk/Internal-Non-Raid-Controller-Desktop-Bracket/dp/B07ST9CPND/ref=sr_1_1?dchild=1&keywords=SI-PEX40139&qid=1609347892&s=computers&sr=1-1

Bad RAM and unclean shutdowns are the most common reasons for filesystem corruption.

CRC errors don't reset, as long as you don't get more it's fine.

That's very bad, I usually get 99% plus, even on very damaged disks, data on that disk will be mostly garbage, possibly with an unfixable filesystem, and it will create the same amount of corruption on the other rebuilt disk, not sure if it's worth it but I can still post the instructions if you want.

JMB585 chipset is good, but not all controllers are the same, it might be bad quality, or just a problem with that one, this is a good quality JMB585 controller: https://www.sybausa.com/index.php?route=product/product&path=64_181_85&product_id=1027

Use -L.