JorgeB

BTW you can find out the actual device with: dmsetup info /dev/dm-X If it's an array device it will show md1 for disk1, md2 for disk2, etc, if it's a pool device it will show sdX1 or nvmeXn1p1, then just check the main page to find out which pool it belongs to.

This is because you're using encryption, they are in from dm-0, which is disk1, dm-1 is disk-2 and so on, also just noticed that at one time you disk14 wasn't mounting, so cache was dm-13, after that disk14 was mounting so cache is now dm-14.

dm-13 is the cache device (/mnt/cache) dm-14 is the ssd pool device (/mnt/ssd)

Nope, just all plugins aren't loaded, alternatively you can uninstall just the UD plugin and reboot.

Disk3 is failing and needs to be replaced.

Run it again without -n or nothing will be done.

Please post the diagnostics.

That looks fine, now post new diags after array start in normal mode.

Start the array in maintenance mode and post the output of: xfs_repair -v /dev/md3

Google TB vs TiB, GUI shows TB.

This is usually hardware related, it can usually be suppressed with this: https://forums.unraid.net/topic/111161-pcie-errors/?do=findComment&comment=1013378

Feb 3 11:22:12 HarryPotter2 kernel: ata2.00: status: { DRDY } Feb 3 11:22:12 HarryPotter2 kernel: ata2: hard resetting link Feb 3 11:22:13 HarryPotter2 kernel: ata2: SATA link down (SStatus 0 SControl 300) Feb 3 11:22:13 HarryPotter2 kernel: ata2: hard resetting link Feb 3 11:22:22 HarryPotter2 kernel: ata1: softreset failed (1st FIS failed) Feb 3 11:22:22 HarryPotter2 kernel: ata1: hard resetting link Feb 3 11:22:23 HarryPotter2 kernel: ata2: softreset failed (1st FIS failed) Feb 3 11:22:23 HarryPotter2 kernel: ata2: hard resetting link Feb 3 11:22:24 HarryPotter2 kernel: ata2: SATA link up 6.0 Gbps (SStatus 133 SControl 300) Feb 3 11:22:24 HarryPotter2 kernel: ata2.00: configured for UDMA/133 Feb 3 11:22:24 HarryPotter2 kernel: ata2: EH complete ... Feb 3 11:25:26 HarryPotter2 kernel: ata1: SATA link down (SStatus 0 SControl 310) Feb 3 11:25:26 HarryPotter2 kernel: ata1: hard resetting link Feb 3 11:25:36 HarryPotter2 kernel: ata1: softreset failed (1st FIS failed) Feb 3 11:25:36 HarryPotter2 kernel: ata1: hard resetting link Feb 3 11:25:37 HarryPotter2 kernel: ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310) Feb 3 11:25:37 HarryPotter2 kernel: ata1.00: configured for UDMA/33 Feb 3 11:25:37 HarryPotter2 kernel: ata1: EH complete Feb 3 11:27:53 HarryPotter2 kernel: ata43.00: exception Emask 0x10 SAct 0x0 SErr 0x4010000 action 0xe frozen Feb 3 11:27:53 HarryPotter2 kernel: ata43.00: irq_stat 0x80400040, connection status changed Feb 3 11:27:53 HarryPotter2 kernel: ata43: SError: { PHYRdyChg DevExch } Feb 3 11:27:53 HarryPotter2 kernel: ata43.00: failed command: READ DMA EXT Feb 3 11:27:53 HarryPotter2 kernel: ata43.00: cmd 25/00:40:08:c5:55/00:05:00:00:00/e0 tag 23 dma 688128 in Feb 3 11:27:53 HarryPotter2 kernel: res 50/00:00:07:c5:55/00:00:00:00:00/e0 Emask 0x10 (ATA bus error) Feb 3 11:27:53 HarryPotter2 kernel: ata43.00: status: { DRDY } Feb 3 11:27:53 HarryPotter2 kernel: ata43: hard resetting link Feb 3 11:27:58 HarryPotter2 kernel: ata43: link is slow to respond, please be patient (ready=0) Feb 3 11:28:03 HarryPotter2 kernel: ata43: COMRESET failed (errno=-16) Feb 3 11:28:03 HarryPotter2 kernel: ata43: hard resetting link Feb 3 11:28:07 HarryPotter2 kernel: ata43: SATA link up 6.0 Gbps (SStatus 133 SControl 310) Feb 3 11:28:08 HarryPotter2 kernel: ata43.00: configured for UDMA/33 Feb 3 11:28:08 HarryPotter2 kernel: ata43: EH complete Constant errors in multiple disks, this is usually a power/connection problem.

No, to see which drive it is.

Not a big deal, it's just a warning, it appears to be caused by the UD plugin, reboot in safe mode if there's no more warning you can post in the UD support thread to see if anything can be done about it.

PCIe 2.0 models will also work and still have enough bandwidth for 8 disks, so if you can get one much cheaper it's a valid option. Yes, if it's flashed to LSI IT mode like mentioned in the listing. That's and older PCIe 2.0 model, but still a good option if it's cheaper.

You can let Unraid emulated the disk, if it's not yet disabled, then just replace it when the new one arrives.

Please post the diagnostics instead.

Assuming the emulated disk is mounting and contents look correct: https://wiki.unraid.net/Manual/Storage_Management#Rebuilding_a_drive_onto_itself

I never used it but as I understand Ipvlan should work the same as before for connectivy.

Yeah, they're not cheap, but sometimes you get what you pay for.

Removing a drive from a pool then adding one later is fine, you just can't do a direct replacement, that's the broken part.

The one I mentioned is available in Europe, got a few from Amazon.es, nobrand direct from China might be OK, might not, once got one, also an Asmedia, that kept giving CRC errors.

Yes, though it's not just the chip that's important, rest of the controller might be garbage, try to stick with known quality models, e,g,: https://www.sybausa.com/index.php?route=product/product&product_id=151

That looks normal to me for 1 year, previous Unraid releases had a problem with excessive writes, IIRC one of my cache SSDs at some point was writing about 3TB/Day, most of the reason it has now close to 1PB total writes.

Around that yes, you can easily calculate based on this attribute: 241 Total_LBAs_Written -O--CK 099 099 000 - 101015376296 101015376296 x 512 (sector size) = 51 719 872 663 552 bytes, or 47.04 TiB