Vr2Io

Call trace have found in previous log, Do you perform memory test to prove system stability ?
 
 
Yes .... some post say newer Intel platform trend unstable with current Unraid, but I can't verify that because I stay in 9th gen.

This likely existing USB stick haven't legacy boot sector / file. Simple run makebootable.bat ( something like that ) at USB stick root folder ( if under Windows ) will solve the problem.

As you have RAID-Z2 pool, so it allow two disk failure / missing, and you have 12TB data which can't fit in one 10TB disk.
 
You shouldn't destroy the RAID-Z2 pool, just cleanup two disk under Unraid and format it, then boot back TrueNAS and confirm it can mount both, then copy all data to those disks.

My main storage build are x299 9800x, main reason was it support 8 memory slot ( 32GB x8 ), I setting its power limit PL2 at 66W, so clock frequency maintain to ~3.3Ghz, so far so good.
It have two ZFS pool, 8 disk raidz1 and 12 disk raidz1x2.
 
If you not need much PCIe lane and memory, then no reason to go in those platform. My 2nd and 3rd build ( not storage purpose ) are 9700k and 5600h.

Base on that, it doesn't like UPS nor battery have problem, suggest turn off APCUPSD and install NUT for troubleshoot why APCUPSD will interpret the battery were fault and shutdown immediate. ( NUT can setting don't shutdown, so you can further troubleshoot ** pls stop array for safe ** )
 
After you disable all UPS monitoring, you can in command prompt type APCTEST to set battery date or perform selftest.
 
And you need fix why NUMXFERS so often !!!  ( Of cause, it may UPS fault )

Add the device to HA docker. ( I only try HA docker to handle the dongle in early testing and I use Zigbee2mqtt currently, so it may have some different, anyway almost that way. Or you may ref. below youtube clip )
 

 
Depends on which type Sonoff dongle, integrations / config it with correct protocol ( ZNP / EZSP ).
 

 

 

 
 
 
 

There are some rack product have SAS / SATA, so no mess data and power cable. And price also affordable.
 
For example, Gooxi RMC2125 just need usd520. ( Mfg price)
 

 
 
Almost same even little bit more.

Check DNS first, likely DNS resolve issue.

只要沒有 parity 2, 可隨便轉換盤位.

If error come from sdc and not both with sdh ( another M600 in cache pool ), it likely sdc were bad.

因為區塊 trim 後, 重讀這些區塊可以係不確定, 這樣會造成  parity 驗證不正確, 但有啲ssd會係 trim 後讀出全零, 這樣便沒有問題. 因此 Unraid 直接不允許在 array disk trim. 其實 trim 不 trim, 好多時都係冇大問題, 不用糾結.

Unraid OS usually work well with different hardware, just set not auto start array, backup USB stick by build-in backup function, then booting in new hardware and check does all storage device correctly detect and go on.
 
I further check  i7-4790S, it have iGPU and quicksync, would you really need Nvidia GPU ?
 
The other solution is buy a new 8 series mobo and give 2nd life for i7-4790S ( China have those 8 series new mobo very cheap ( less then 50usd ), dual LAN and support NVMe booting ( although you don't need that for Unraid ), build to ITX and SFX PSU and make it as small footprint as possible. I have buy one for a friend too, so far so good. ( But wrong BIOS setting will need CMOS reset )
 
You don't need 600w PSU, my 9700K build just use a 300w SFX PSU.
 

 
Or This one in Amazon ( but not support PCIe NVMe and single LAN ), size are DTX ( between ITX and mATX ), Cooler master NR200P support this mobo and max with two 3.5HDD.
https://www.amazon.com/MACHINIST-Motherboard-H81-Processor-H81M-PRO/dp/B09DVCV3LT
 

 

md_write_method="auto"
 
Try enable reconstruction write first.
 
Even that, it seems a bit slow, less then 30MB/s, but this also depends on large / small file size.

Seagete NVMe ?  if the pool are single storage pool, you can safe add it back.
 
But I also aware Seagate NVMe at UD haven't show the mount button, this may indicate it have file system problem.

You can add them in your ignore list, but can't add Moderators.

If IPv6 and docker host access not need, pls try disable both and check problem solve or not. Pls also use IPVLAN too.

No expander have detect, check the power , SAS cable, LED status, and touch heatsink warm or not.
 
EDIT: You got no SAS cable between controller.

Pls also run memtest, your system show some odd reading, configured in 1280MT/s or 257MT/s.
 
My system show FYR.
 

 
 
Memory Device Array Handle: 0x000C Error Information Handle: Not Provided Total Width: 64 bits Data Width: 64 bits Size: 4 GB Form Factor: DIMM Set: None Locator: A1_DIMM0 Bank Locator: A1_BANK0 Type: DDR3 Type Detail: Synchronous Unbuffered (Unregistered) Speed: 667 MT/s Manufacturer: Kingston Serial Number: BA19045C Asset Tag: A1_AssetTagNum0 Part Number: KHX1600C9D3/4GX Rank: 2 Configured Memory Speed: 1280 MT/s Memory Device Array Handle: 0x000C Error Information Handle: Not Provided Total Width: 64 bits Data Width: 64 bits Size: 8 GB Form Factor: DIMM Set: None Locator: A1_DIMM1 Bank Locator: A1_BANK1 Type: DDR3 Type Detail: Synchronous Unbuffered (Unregistered) Speed: 800 MT/s Manufacturer: Undefined Serial Number: 00001761 Asset Tag: A1_AssetTagNum1 Part Number: TIMETEC-UD3-1600 Rank: 2 Configured Memory Speed: 257 MT/s  

Yes it is a plugin, SimonF version support all Unraind version.
 
 

Because docker network have IPVLAN and MACVLAN mode, in IPVLAN mode, all containers use same MAC address ( real interfaces's MAC address ). Most consumer network equipment will list / count device by MAC address ones only, so you will got either one of IP's with same MAC address instead same MAC address with different IP.

03:00.0 Ethernet controller [0200]: Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller [10ec:8168] (rev 06)
    Subsystem: Gigabyte Technology Co., Ltd Onboard Ethernet [1458:e000]
    Kernel modules: r8169
04:00.0 Ethernet controller [0200]: Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller [10ec:8168] (rev 06)
    Subsystem: Gigabyte Technology Co., Ltd Onboard Ethernet [1458:e000]
    Kernel modules: r8169
 
You have blacklist the driver ?
 

 
 

 
 
 
 

This is docker image file usage.

In Web term try press F5 to refresh.

When in virtio NIC, the link speed just a number, it can perform more then it. ( In some physical Switch / NIC also that, i.e. RJ45-SFP+ media converter, throughput could larger/smaller then link-speed ) 
 
Someone bridge NIC in Unraid network and you may test any performance different by Windows VM. I haven't bridge NIC.

Yes all lost, in fact no matter which storage technic, it can crash/corrupt, so backup was important. Raid0 just for speed, on store and recovery.
 
Btrfs RAID0 no need all disk in same size.