navigat0

Yes,I did address that thread in my initial post. While not being an issue previously I have completely disabled c-states. I will try to flash a newer BIOS, but as you've suggested nothing suggests a hardware issue, so I'm not putting a lot of weight in that solving my problem here. I'm less than thrilled to see nothing useful in the syslog prior to the system becoming unresponsive. That being said, it has been running since your post, with downtime associated to me re-installing the RMA'd RAM.

Well that didn't last very long. Nothing useful in the syslog, but I attached it anyway. I've included diagnostics too, and hopefully something stands out because I'm at a loss and until the upgrade this thing was rock solid (even, apparently, with a failing DIMM). SyslogCatchAll-2023-09-02.txt SyslogCatchAll-2023-08-31.txt SyslogCatchAll-2023-09-01.txt tower-diagnostics-20230903-2005.zip

Odd. Made the change again. S L O W E R this time, revisiting the pages after each change, and the service did start. I'll start the clock on uptime again and report if another failure occurs.

I did try this, but maybe I wasn't properly prepared for the change because after changing the setting the docker service would not start. I can try again. Are there any specific steps other than changing the docker option in setting with the docker service stopped?

A few weeks ago I upgraded from 6.9 to 6.13. No other changes were made prior to this issue beginning. Since then my server has suffered irregularly timed unresponsiveness. Sometimes a few hours, sometimes many days pass by before the system goes unresponsive and requires a reset/ hard boot. I've set up remote syslog and captured a number of instances but thus far have been unable to determine a root cause. I've read a number of similar posts from over the year, but many don't have any final actions taken or resolution. I've closely followed the steps outlined here regarding AMD setups: While c-states were set to auto and had no ill-effects running on 6.9 for years, I have completely disabled them. Memtest determined that one of my DIMMs was in fact bad, but I've removed that DIMM and am currently experiencing the issue running on the single remaining DIMM which has passed multiple tests over a period of days. As with any troubleshooting scenario I've been slow to make changes and only change one thing at a time to determine the true impact. My next step might be to update the BIOS, but again, that one of those things that you don't really do without good reason - security patch, new functionality, or big fix. Additionally stepping back to an older version of Unraid is always a possibility as well. Assistance resolving this issue is appreciated. Board, CPU and RAM are call certified compatible https://pcpartpicker.com/b/qx7WGX SyslogCatchAll-2023-08-30.txt tower-diagnostics-20230824-1625.zip

Just curious as I recently had a flash drive failure as well and maybe I'm missing something. I was using the flash backup to unraid.net and I successfully downloaded my flash backup file, created a new flash boot on the replacement drive, and ran into the same key error above . I'd already copied the downloaded key from my unraid.net my servers instance (same place I downloaded my flash backup) to the flash drive's config folder (BTW this step is not well documented) but had to go through the support option to generate a new key for the replacement drive. So my question is, what's the point of backing up the key if you can't restore it to a new drive to use without having to contact support or request a new key? And more importantly, that being the case, what if you suffer more than one USB flash boot failure in a year (flash drives aren't the most reliable storage and I've been using name brands)? Are you expected to purchase a new key?

I'm creating a new Windows 10 VM and following the same methods (Spaceinvader1 and Proxmox) that have worked previously for me, but this time I'm stuck. I've read countless posts on this forum and none of them seem to fit my situation exactly but did provide some good insight on things to try. Config and logs are attached - this is running in unraid 6.8.3. I've created the VM as I normally do preferring OVMF to allow pass-through of various devices (My existing Win10 vms are all OVMF). I'm using a Win ISO downloaded yesterday, and a recent copy of the VirtIO drivers. I booted fine to the WinISO, and when it came time to choose an installation location I loaded the virtiostor driver to access the vdisk. I selected the disk and Windows installer went on it's merry way installing. I left expecting to come back to the "first-run" setup for Windows but the VM is stuck booting to the UEFI shell. Reset goes through the BIOS startup and boot to DVD option, then back to the UEFI shell Continue goes to boot to DVD option and then to the UEFI shell You can see the misc device appears to be the vdisk image yet it won't boot. If I boot to the ISO and use the system repair tools -- Running command prompt and diskpart list-disk results in "There are no fixed disks to show." Running command prompt and browsing the local filesystems I actually see on drive X: what appears to be my windows installation (program files, users, windows, etc.) Startup Repair couldn't repair your PC. W10-Template-config.txtW10-Template-log.txt Thanks in advance to all!

Still no luck and sorry for the long-ish post I tried the approach testdasi posted above, by modifying the controller to be a virtio-scsi passthru controller (even re-addressing it to match my optical drive's lsscsi address and swapping it with my network interface). Windows does indeed install the drivers for the passthru controller and assigns the optical drive appropriately and as usual all devices are working properly in windows but sadly media is still not recognized. Additionally I decided to test another (slightly older) optical drive connected via a SATA/IDE -> USB conversion kit. Immediately the Windows VM sees the drive and it can actually read/write optical discs of course it's severely limited by the USB 2.0 bus. I'll verify the SATA drive when I get a moment to pull it from the tower. At this point I can only assume either the drive itself is bad, the onboard SATA controller doesn't like the device or something in the middle-layer of unRAID is causing an issue as the drive is properly recognized but the data layer is where things fall apart. On another sad, but related, side-note (and likely for one of the same unknown reasons listed above) my MakeMKV and Ripper dockers can both see the drive, but never identify any media in the drive. If I go the USB route with the IDE drive for the dockers, Ripper will identify media is present, but it's always a data disc (regardless of the actual media (DVD, CDFS, or data) and it never actually creates a proper image and locks the device up. I've set up a lot of complicated things in unRAID as a relative noob to the platform including migrating a Win10 ESXi VM with hardware GPU passthru, setting up a reverse-proxy, deploying OpenVPN-AS with client-cert auth, and yet, optical media management seems to be taking (or wasting) the most time. I'm hoping that continued challenges with optical media and unRAID prompt the development of direct access support (Unassigned devices maybe?) or fully supported passthrough. In the meantime if anyone has an extra PCIe SATA controller just layin around they'd like to donate DM me I hope all of my efforts can help someone else succeed or save them the trouble.

A little more info on my setup. The DVD drive is currently attached to the on-board SATA bus and not a separate controller. The onboard SATA is also being used for all of my unraid drives. That' being said, the DVD drive is visible in the VM and device manager reports that the drive, the virtio SCSI, and virtio SCSI passthrough devices are all working properly. Looking at the address, children, and parent attributes of these controllers and my drive it looks like the VM sees the drive attached to the virtio SCSI passthrough controller as expected. I've tried changing the device controller ID to another number but that results in the VM not seeing the drive at all, and the creation of additional LSI SCSI controllers that just show up to windows as unidentified SCSI controllers for which no drivers are available. I've also tried adding an alias field (hostdev0) and the rawio set to yes, but in these cases the drive behaves exactly the same - windows sees it, but never recognizes any disc as valid and asks for me to insert a disk. Being there are no issues the Windows event log is clean. Here's the relevant lines from the VM log (no errors) and it appears the drive/ device is mapping properly, but the media data is not usable. -device virtio-scsi-pci,id=scsi0,bus=pci.0,addr=0x9 \ (Windows sees this as the virtio SCSI pas-through controller based on the address and based on the children attribute the DVD drive is listed as a drive under this controller) -device lsi,id=scsi1,bus=pci.0,addr=0xa \ -device lsi,id=scsi2,bus=pci.0,addr=0xb \ -device lsi,id=scsi3,bus=pci.0,addr=0xc \ (These were created when changing the controller #. They have been removed from the config and no longer show in my current log) -device virtio-blk-pci,scsi=off,bus=pci.0,addr=0x6,drive=drive-virtio-disk3,id=virtio-disk3,write-cache=on \ (This is the virtio SCSI controller to which the small 2nd vdisk is attached) -drive file=/dev/sg3,if=none,format=raw,id=drive-hostdev0 \ -device scsi-generic,bus=scsi0.0,channel=0,scsi-id=0,lun=0,drive=drive-hostdev0,id=hostdev0 \ (The device in question currently configured as the following) <hostdev mode='subsystem' type='scsi' managed='no' rawio='yes'> <source> <adapter name='scsi_host3'/> <address bus='0' target='0' unit='0'/> </source> <alias name='hostdev0'/> <address type='drive' controller='0' bus='0' target='0' unit='0'/> </hostdev> I don't have any other SATA optical drives nor do I have a SATA PCIe controller to test with.

Running unraid 6.7.2 on AMD https://pcpartpicker.com/builds/edit/?userbuild=qx7WGX Using the lsscsi command I was able to identify the drive in the unraid console - it's an LG DVD writer connected via SATA to my MB ... [3:0:0:0] cd/dvd HL-DT-ST DVDRRW GSA-H30L S856 /dev/sr0 ... I'm able to pass through the DVD writer using <hostdev mode='subsystem' type='scsi' managed='no'> <source> <adapter name='scsi_host3'/> <address bus='0' target='0' unit='0'/> </source> <alias name='hostdev0'/> <address type='drive' controller='0' bus='0' target='0' unit='0'/> </hostdev> After starting the Win7 VM and installing the SCSI virtio SCSI device driver the drive is recognized and device manager says the device is working properly, but doesn't read any media inserted, only telling me to "Insert a disc into drive D:" This is the case with Audio, Data and DVDs. If added to my Win10 VM there is no compatible virtio SCSI driver for Win10 and without it the drive cannot be recognized or used. Any thoughts?