duffbeer

I ended up rolling back to 7.2.3 and the VM loaded fine.

I updated unRAID last week and didn't think to try my VMs till today. I had a template "Windows 10 Gaming 3" (that is actually running Windows 11 now) which, when I attempted to boot today, only gave me a black screen. After several forced shutdowns via the VM Manager, to troubleshoot, I removed the NVIDIA passthru card and used VNC. I get the Windows System Recovery terminal, and Command Line for DISKPART shows no disks or volumes are present. I'm using a WD NVMe as passthru (no issue before the upgrade) I created a new template (specifying Win11 this time) and set the same settings. Still the same result. Do I need to downgrade back to 7.2.3 or is there maybe something I don't know about that can fix this? ChatGPT kept calling out this as the probable issue: 2026-06-03T15:53:36.813915Z qemu-system-x86_64: vfio_listener_valid_section received unaligned region 0000:06:00.0 BAR 4 mmaps[0] iova=0x847fbf00 offset_within_region=0x0 qemu_real_host_page_size=0x1000 2026-06-03T15:54:19.062102Z qemu-system-x86_64: vfio_listener_valid_section received unaligned region 0000:06:00.0 BAR 4 mmaps[0] iova=0x847fbf00 offset_within_region=0x0 qemu_real_host_page_size=0x1000 2026-06-03T15:54:20.119240Z qemu-system-x86_64: vfio_listener_valid_section received unaligned region 0000:06:00.0 BAR 4 mmaps[0] iova=0x847fbf00 offset_within_region=0x0 qemu_real_host_page_size=0x1000 PCIe ACS Override setting is enabled, btw. hagrid-diagnostics-20260603_1111.zip

That would be a message of Errors found during Parity, right? Will Parity correct an error found on disk, in the case of bitrot or something?

Well crap. Just "bitrot" then, maybe?

hagrid-diagnostics-20260528-1128.zip

Well, I got the green PASS banner, so I assume my RAM is good. Any other thoughts?

Alright, I'll start with memtest. Parity Check only found errors back in Feb, the first time since 2022, and it was because of a bad HSA board, which has now been replaced.

I've had family reports of movies disappearing from the server, things they know they've watched. I didn't believe them at first, but noticed in Tautulli that there was an entry for a watched movie that was missing from the disk. I installed Dynamix File Integrity and have let it run once a month for the past several months. I've had a few files on disk show up as corrupted now. When I go to the files (not movies in the latest examples), the filename is changed to something with the same file extension, but with a tilde in the filename, like it's an old deleted Windows file. I have Parity run once a month, and a few days before that I have Dynamix File Integrity run. What can I do, or how can I find the root cause of this file corruption, and prevent it from happening?

Thanks. I had it set as DHCP Reserved at the router, but didn't set it statically in unRAID. I'll give that a try.

My server is not exposed to the internet, at least not for SSH access. Usually I have an active SSH session with the server. When I let it idle for a day or two and I try to re-esatblish a fresh session from inside the network, I get rejected "ssh: connect to host hagrid.local port 22: Connection refused" I have to go into Settings > Management Access > and toggle SSH from Yes to No, then save, and toggle it back. After that I can SSH again. I just rebooted the server, attempted to SSH in to look at some logs, and was rejected. I didn't do the toggle dance to get SSH enabled again before collecting the diagnostics. Any idea what is happening? I see this in the syslog: Jan 16 09:13:55 Hagrid sshd[21869]: error: Bind to port 22 on 192.168.1.25 failed: Cannot assign requested address. Jan 16 09:13:55 Hagrid sshd[21869]: fatal: Cannot bind any address. hagrid-diagnostics-20260116-0914.zip

Had same issue as you. Logged into Proton and saw that the WireGuard config had expired (I didn't know it would do that!) Enabled "NAT-PMP (Port Forwarding) Enabled "VPN Accelerator" Chose a server that is "P2P friendly" Generated a new config Backed up the old config on the host and replaced it with the new one. Interface now loads. Hope it's the same for you.

Summary of the results: Took the 10TB drive out and hooked it up to a write-blocker and put a 14TB drive in a USB sled, then used DDRESCUE to clone the 10TB drive to the 14TB drive, which I ran UFSExplorer against it I bought UFSExplorer and scanned the 14TB with the 10TB ddrescue'd image (let it run all night). I like that it saved the results in a database that I could load later and not have to rescan the drive again. I put in the second 14TB drives and let Unraid "rebuild" the 10TB drive, which didn't help in the end because it was "unmountable". I did a parity swap with the 14TB drive and the original 12TB parity drive. I ended up doing a New Config and transferring the files since the files were pretty much all there on the drive, just the directory structure got corrupted. I recognized the files based on their names so I manually exported them from my write-blocker attached drive on my Desktop computer over the network to the correct shares I used Unbalance to move all the files on the 4TB drive to the 12TB drive, then I replaced the 4TB with the additional 14TB drive I had and let the array rebuild. That worked just fine. After that the new 14TB drive got over 1.4m UDMA CRC errors in a few hours, so I ended up replacing my (SI-PEX40137) Syba 8 port expansion card SATA cables, but it didn't help. I kept getting UDMA errors. I ended up replacing it with a LSI card from The Art of Server on ebay and so far so good. Thanks for the help everyone!

I'm back! Got two new 14 TB drives. Using UFSExplorer, I'm able to recover just about everything from the 10TB drive. I've put 14TB drive "A" into my array (drive "B" is hosting a DD image of the failed 10TB drive externally, and won't come into play for a while) I set the 14TB drive as the parity drive and set the old parity drive (12TB) as a data disk. It is copying over the old parity (which appears to have been corrupted since the Emulated drive wouldn't load anything) Do I bother with the Parity Copy? It's going on now, but seems like a waste. Can I cancel it, copy over the recovered files from the 10TB drive, and THEN rebuild parity?

Yes, sorry, I rebuilt disk5 from parity.

Rebuild meaning rebuild parity on the same disk? Yes Backups? I'm not sure what files were on this disk, so I'm not sure :( My thoughts now are to take out this corrupted 10TB drive, hook it up to a write blocker and dd it to one of the 14 TB drives. I'll put the other 14TB drive in it's place in the server and make it the new parity drive. I'll try to carve files from where I will dd the drive and output them into the array. Thoughts?