testdasi

You need to copy-paste your VM xml here. (when copy-pasting text from Unraid, please use the forum code functionality - the </> button next to the smiley button - so the text is formatted correctly). Also describe exactly what you meant by "loops", preferably with screenshots. While you are at it, would be useful to attach diagnostics to any query. Tools -> Diagnostics -> attach zip file.

No. You need at least 1 device in the array. It could even be a USB stick (which is NOT your Unraid USB stick) but there needs to be something. The only reason that would require an additional SSD for cache (regardless if SATA or another NVMe), instead of just using the 970 EVO for cache, is that you pass the 970 Evo through as a PCIe device to a VM (i.e. exclusive use of the 970 Evo by the VM). PCIe pass through would maximise performance and allow trim (as compared to ata-id pass through method - which you should not use with NVMe). It's not that putting vdisk on the 970 Evo would be terrible. Performance would still be good and should still be better than a SATA SSD. It's just that it wouldn't be "max". So instead of buying an additional SSD for cache, I would suggest you just stick to a SINGLE 970 Evo in the cache pool. (Note: there were reports of poor performance with multi-drive btrfs cache pool and it seems to be correlated with Samsung unusual TLC block size. Even though I didn't have such issue, it probably is prudent to avoid multi-drive cache pool with Samsung SSD for now).

9p performance (i.e. mounting a share through virtio) is terrible and it has been like that for quite some time. Theoretically it is possible to tune it to make it better but that has never been proven in practice. And no, there is nothing wrong at all to have share specific to a VM (or multiple VM's) being accessed through SMB / NFS. That's the whole point of Unraid user control. Maybe that's what you're missing?

Those are some very old GPU's so it would be quite hard to know what would work without the person actually having the card. You might to ask the "someone" who suggested the 8400 GS on the config.

When your server crashed, did it freeze or crash? Freeze = internal light still on and/or display still showing stuff and/or you need to press and hold power button to hard shutdown before being able to switch on. Crash = light off, no display, a normal press of power button to switch it on. If it's (1) then potentially issue with CPU / mobo / RAM. If it's (2) then potentially power related issue e.g. PSU problem or even a shorted connector. Your syslog does not show any clue at all so it will be hard to diagnose so I can narrow it down to the above. Btw, the mover constantly tries to move your libvirt and docker images. Check your share settings.

Yes you should. As you start to use the gdrive for things beyond media storage (which you will once you discover how convenient this whole gdrive thing is), you will start to desire separating data for easier organisation and/or approach the limits of gdrive (yes, even unlimited has limits) and/or control access. That's where team drives come in. It's better to start using tdrive now then to wait till you need it and then have to wait for things to be moved about. You can double check the encryption just by trying to access the file from the gdrive website. The file names should be jumbled up and if you download the file, you should not be able to make any sense out of it.

Do you have Hyper-V turned on in the VM template? If not, you need to start a new template because turning it on in the GUI sometimes doesn't work. Advanced warning: turning Hyper-V on inside a Windows VM may stop the VM from booting. Been there, done that.

The onboard audio is on the same bus as a USB controller (0a:00.4 and 0a:00.3) so perhaps try passing through both. Make sure to check which controller your USB stick is on first. You should never pass through the controller which the Unraid USB stick is on - if you need to do so, you have to shutdown, put the Unraid stick on a different port and start. To check which port is on which controller, you need the VFIO-PCI Config plugin (from the app store) and a spare USB stick. Then Settings -> VFIO-PCI.CFG and it will take you to the page that kinda looks like your System Devices page but under the USB controller it will tell you which device is plugged in to it. You then use the spare USB stick and plug in and then refresh the page and the stick will then show up under a certain controller - voila.

Up to you. If you set everything to autostart, including array, the server can boot in a few minutes or so. You can also install Putty (google it) and set up a shortcut such that it would automatically connect and send a shutdown console command to the server which will shut it down. That way you can shut the server down by a double click and then shut down your main PC. You can even go one level up and use the User Scripts plugin and a regularly scheduled bash script (let's say every 30 minutes) that try to detect a file from your main work PC. Once your main work PC shutdown, the file would not be accessible and then the script would shut down the server. It's relatively quite advance. There are also ways with Linux to deal with the power button (think it's in /etc/acpi/acpi_handler.sh) so pressing the button will send shutdown command instead but that is super advance Linux stuff that I have no idea how it would work. Or you could just leave it running 24/7 if power cost is not too high. I leave mine running 24/7 - used to think it's a waste of electricity until I discovered folding at home so I leave my server "idling" at 65 degrees 24/7 for the advancement (and right now the survival) of the human race.

What you are proposing would be terrible for what you are trying to do. If something is wrong with your source, the wrong-ness will automatically cascade down to all backups. That would defeat the whole point of having a backup. In addition, what you are trying to do would just be creating more versions of the same online backup. So if let's say there's a power surge, all your data is destroyed anyway, regardless of how many copies you have. When dealing with vital data, you need to think of it in "Four-O": Original copy Online backup - a medium that is plugged in and connected with the original copy most of the time e.g. an array dup of data Offline backup - a medium that is disconnected and unplugged except during data transfer e.g. an external USB drive Offsite backup - a medium that is not in the same physical premise as the original copy e.g. cloud backup or even an external drive at a relative's place So instead of propagating many versions of the same copy i.e. making many online backups, you should, for example, take 1 HDD out and use it externally. If the data is that critical to you, to the extent that you want it mirrored across many drives, I would assume you would want the full Four-O.

You can run multiple defrags at the same time as long as you don't have parity (with the caveat that you won't know which thread will be used for each defrag instance). With parity, running multiple simultaneous defrag will slow it down exponentially.

Run a diskcheck of your USB stick on a Windows machine and rebuild it. Could be a corrupted stick.

Either iGPU or onboard would be fine as long as the mobo BIOS let you boot Unraid with it. You can put the 860 Evo to cache and mount the Intel as unassigned device (install the Unassigned Devices and Unassigned Devices Plug plugin from the app store). With regards to VM fig, have a read of the VM FAQ, it will answer a lot of your questions. E.g. below is link to the post about how to enable trim on vdisk

Did you install the driver using the download from Nvidia website or did you use the default Windows driver? I think it's the default Windows driver which might be why you can get a display without the sound device. It's still rather peculiar that it would even work though. Anyway with regards to your xml, you probably will need a vbios to get both GPU and HDMI audio to work (with proper Nvidia driver) since it's the only GPU in your system. SpaceInvader One has a tutorial on how to get vbios so perhaps watch that.

With 6 drives, having dual parity is an overkill based on Backblaze HDD failure stats. You have to be super risk averse and/or store super important data on the array to require that level of protection. I haven't heard of RC 580 graphics card. I think you meant AMD RX 580? If that's the case, the very first thing you need to do is to try booting Unraid with the mobo integrated ASMedia GPU. The only success stories of passing through RX 580 on here involve booting Unraid with another GPU. Since both SSD's are SATA, using either of them exclusively for a VM (i.e. passing through using ata-id method) will not yield you much (if any) perceivable diff, as compared to putting a vdisk on the drive. With vdisk (and appropriate config), Windows will detect the vdisk correctly as thin-provision and thus enable trim. An ata-id passed through device doesn't support trim, as far as I have managed to try in the past. There is the scsi-bus pass through method which may (or may not) enable trim but I haven't tried that when I had a spare SATA SSD. So if you really want to give the VM exclusive use of an SSD, you are better off with Option 1. That is assuming, being an enterprise MLC SSD, the S3500 runs better without trim.

Create a Windows VM and mount those files in VM. I am not aware of anyway to direct mount vhdx in Unraid.

As a starting point, attach your Diagnostics (Tools -> Diagnostics -> attach zip file). Then copy-paste the PCI devices section from Tools -> System Devices + attach your current xml. When copy-pasting from Unraid, use the forum code functionality (the </> button next to the smiley button) so the code is formatted correctly. There's no need to manually section things out like your manual post since the code block will be obvious. To be honest, this is a rather peculiar case. I have a GT710 and it would not work without the HDMI Audio device passed through. In your VM what works with the GPU without HDMI Audio, what driver do you use?

No there isn't.

Spontaneous downgrade is NOT possible. You might want to think very carefully about what happened because if you can eliminate all possibilities of user error then there is only one possibility left - you were hacked.

One thing I forgot to caveat. I was talking in terms of Unraid core experience. If you plan to ever use ZFS (e.g. ZFS plugin "app" that is not officially supported by Unraid i.e. outside of the Unraid core experience) then ECC is considered a must-have requirement. I do agree with your point above. When picking between spending on a backup vs ECC RAM (for Unraid uses), I would pick a backup every time.

The point about ECC RAM being slower than non-ECC RAM is not relevant since you shouldn't be overclocking RAM on Unraid. Those advertised high speed and fast timing are certified overclock and is an overclock nonetheless. ECC won't prevent system crash or data corruption in the sense that if you use it, you won't have those things taking place. It only protects you against crash / corruption in a very specific case that is a single-bit error. Whether that matters or not (vs the cost of ECC RAM) is entirely personal preference. I have run both ECC and non-ECC and my personal anecdotal experience is that it makes no perceivable diff in terms of stability or data corruption.

Removing it from syslinux will only stop it from appearing on the Other PCI Device section of the VM template. Your existing VM config would still have it and thus when you start the VM, it would still pass through. This is apparent in your latest xml diff, which contains the config to pass through 03:00.0. If you are comfortable with deleting the xml section then do it manually via xml edit. Otherwise, I would suggest starting a new template from scratch (perhaps copy-paste the xml here as well so everyone is on the same page with regards to which VM you are testing).

Unraid freezes when you passed through 28:00.3 probably because the Unraid USB stick is on it. I provided instruction on how to identify what port is on which controller, which I don't think you have followed. Your problem is kinda unusual so logically the cause should be something uncommon. And the one thing that is uncommon about your config is that you passed through the USB 3.1 controller, which nobody has ever successfully done so - at least none that I have seen on here. 100% load on 1 core and being stuck on Tianocore screen CAN be the symptoms the controller failure to handshake with the USB devices. And a non-resetting / problematic controller will fail to handshake with USB devices. So 03:00.0 is the primary candidate of your issue but your refusal to remove it from the config is not helping at all. Remove it from your config and see if the problem still occurs.

Welcome to Unraid. Good job with the cable there. That is very neat. A few unsolicited tips for you: Ask yourself whether the data in cache pool needs RAID-1 or not. If you are not storing any important data on there, you can make your SSD last even longer by separating write heavy and ready heavy data to 2 SSD instead. The write heavy SSD (e.g. for download temps) should be almost empty most of the time except when data is being written to it. That will minimise write magnification and wear leveling and thus enhance your SSD lifespan. Alternatively, the next time you get a new SSD (which presumably should be larger than 128GB, they don't make that size anymore), add that as a single-drive cache pool instead and just cp data over from the old cache pool. Then use the old 128GB for write-heavy stuff. With the right treatment, SSD's last forever (I still have a 128GB SATA II (yes, TWO) that does not have even a single reallocated sector despite going way over its rated endurance rating). Aim to get fewer large capacity drives instead of more smaller ones. The more drives you have, the more likely you will have a failed drive. And make sure to preclear each drive before adding to the array to weed out the "infant mortality". Are you running Unraid Nvidia? If you want to use the 1660 (or 970) outside of a VM for Plex, you will need Unraid Nvidia build. Read up on it (there's a forum topic) before jumping in. Alternatively, you can pass through the GPU to a VM (with an OS that supports Nvidia e.g. Windows) and run Plex in there. There's a lot of overhead with this + passing through GPU isn't the easiest thing to do but that's one alternative to consider. Change your CPU cooler fan config. It's better to set it to push air onto the cooler and then (pulled) through the rear exhaust fan. That way you kinda create a push-pull config that works a bit better than a double pull from the look of it right now.

What do you expect to be your parity drive? At boot, it looks like only 3 HDD were loaded, 2x 4TB Seagate and 1x 1TB WD.