JorgeB

Connections on disk5, shutdown the server, check connections even if it's just removing and re-connecting the cables, power back on and post new diags, disk is still offline on the diags posted.

Unraid has the driver, though IIRC it's not working correctly on v6.8, it should work on v6.9-beta22, please post diags from the beta.

Parity swap procedure needs to be done from start to finish, any other step/interruption will make it stat over, if you're absolutely sure the parity copy completed you could use the invalid slot command, but I would advise just starting over, also 30+ hours is a lot, unless the parity is very large.

Shutting should be perfectly safe. No point, server has two invalid disks that can't be emulated by the single parity.

You're welcome, like mentioned not all SSDs (in fact most of them) are able to sustain 500MB/s writes, 800MB/s means you now have an NVMe device, which is generally considerably faster, though like all devices some models are faster than others.

If it's not that it's likely a problem with the disk.

Thanks for write up, and if you don't mind going to mark this as solved.

You have a lot of stuff (everything?) in the nerdapck plugin installed, make sure you only install what is needed, and of those and IIRC atop is known to fill the log.

Reboot.

All errors appear to have been correct by the scrub, still good to run a second one to confirm, if no more errors filesystem is good now. Correct.

Couple of unrelated observations: -trim is running every hour, that overkill IMHO, once a day is usually what we recommend -CPU is overheating and throttling down, check cooling Jun 20 20:09:11 TazServer kernel: CPU14: Package temperature above threshold, cpu clock throttled (total events = 1) Jun 20 20:09:11 TazServer kernel: CPU15: Package temperature above threshold, cpu clock throttled (total events = 1) Jun 20 20:09:11 TazServer kernel: CPU7: Package temperature above threshold, cpu clock throttled (total events = 1) Jun 20 20:09:11 TazServer kernel: CPU4: Package temperature above threshold, cpu clock throttled (total events = 1) The device error is cause by an unassigned NTFS disk that possibly isn't even no longer connected: Jun 20 20:30:42 TazServer ntfs-3g[12382]: ntfs_attr_pread_i: ntfs_pread failed: Input/output error Jun 20 20:30:42 TazServer ntfs-3g[12382]: Failed to read index block: Input/output error Jun 20 20:30:42 TazServer kernel: Buffer I/O error on dev sdh2, logical block 9852661, async page read Run chkdsk on that disk before using it again.

There's corruption on the cache filesystem, almost certainly caused by 2 of the 3 devices dropping offline in the past: Jun 21 19:45:09 Raza kernel: BTRFS info (device sdc1): bdev /dev/sdc1 errs: wr 574718, rd 407964, flush 6103, corrupt 0, gen 0 Jun 21 19:45:09 Raza kernel: BTRFS info (device sdc1): bdev /dev/sdb1 errs: wr 724802, rd 490716, flush 6103, corrupt 0, gen 0 See here for more info, if a scrub can't correct everything, and it probably won't, you'll need to re-format it.

Yep, since it now includes the most typical options it's not possible to type a custom one, I need to update the FAQ.

Single parity can't emulated 2 disks, disk5 dropped offline so there's no SMART, check connections and post new diags. I assume you still have old disk1 intact?

Seems to be a hardware issue, make sure it's not related to the 3.3v pin problem, google "WD 3.3v pin"

Don't see any drives failing to initialize, to what controller are the missing drives connected?

The main change is that now you can have multiple pools, any pool with more than one device must be btrfs, like always.

Please post the diagnostics: Tools -> Diagnostics

You are transferring at 800Mbits/s, not 800Megabytes/s.

You are confusing gigabit with gigabyte, gigabit LAN can transfer at around 110MB/s, only 10GbE Ethernet can transfer at 1GB/s+

Driver should be fine, try testing with another computer, cable, switch, etc

You can use rsync if it's XFS or rsync/snapshots if it's btrfs.

The command in the FAQ are to use in the balance window on the cache GUI page, if you want to use the console it would be: btrfs balance start -mconvert=raid1c3 -dconvert=raid5 /mnt/cache

Yes, I already asked LT to at least include raid1c3 because it's the best option for raid6 metadata, since it has the same redundancy level. raid1 metadata is fine for raid5, though you could use raid1c3 for increased reliability, raid1c4 is overkill for raid5 IMHO See here how to convert: https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=480421 Note that you can convert just the data profile (dconvert=) or just metadata (mconvert=) by itself, or both at the same time Yes, you can always convert data and/or metadata to any profile, as long the the pool has the minimum required devices.