testdasi

Firstly, you have a very important misunderstanding about NVMe performance. "Hosting" a VM on a RAID-0 pool (by that I assume you mean having vdisk on the RAID-0 pool), by itself, will provide you with virtually no noticeable benefit (other than perhaps shaving a few seconds off boot time - and only "perhaps"). The benefit of NVMe (RAID-0 or not) is heavily workload dependent and so you need to first define what your workload is. For example: Booting a VM only benefit from it once (at boot time) Photo editing is a workload that would benefit from non-QLC NVMe drives (QLC is right now terrible with random I/O) Video editing would benefit from NVMe even QLC (more sequential) Gaming almost doesn't benefit at all from NVMe (as compared to a good SATA SSD). So I would say your effort to get NVMe RAID-0 to work (without first defining your workload) is rather misguided. Yes you will have a bit of bragging rights with benchmarks but is IRL more pain than what it's worth. Secondly, you mentioned "VMs" i.e. plural. Do ALL of your VM's really need NVMe RAID-0 performance? Outside of Linus 7-gamer-1-PC kind of projects, very few users would ever need multiple high-performance VM's (that simultaneously need the benefit of NVMe, let alone NVMe RAID-0). Specifically for Threadripper, there are only 3 general kinds of VM: Workstation VM (more cores for best performance) Gaming VM (only cores from the same die connected to the GPU slot for lowest latency). Miscellaneous VM (mostly accessed remotely for miscellaneous uses such a proxy, router etc.) Of those 3 types, only the 1st type (assuming you have defined the workload as per my first major point) actually would benefit from NVMe. And most users don't need more than 1 workstation VM so why don't you just pass-through the 4 NVMe on the x16 adapter directly to the VM and then use Windows-based striping? Based on a recent test by Puget system (https://www.pugetsystems.com/labs/articles/NVMe-RAID-0-Performance-in-Windows-10-Pro-1369/), Windows 10 striping performance is pretty good, just that you can't boot from it (which is easy since your VM can boot from a vdisk img from the 5th NVMe if that's what you want). Thirdly, and I can't emphasize this enough, RAID-0 is a terrible idea! The pursue of pure (theoretical) speed at the risk of complete data loss is not advisable to 99.99% of users. The 0.01% are mostly tech Youtubers who don't use it as a daily driver.

Not enough details. You need to raise a separate post with question, description and diagnostics.

No it isn't. The 09:00.0 / 1 GT 710 is in Group 14 and the LSI (0a:00.0) is in Group 15. IOMMU group 14:[10de:128b] 09:00.0 VGA compatible controller: NVIDIA Corporation GK208B [GeForce GT 710] (rev a1) [10de:0e0f] 09:00.1 Audio device: NVIDIA Corporation GK208 HDMI/DP Audio Controller (rev a1) IOMMU group 15:[1000:0064] 0a:00.0 Serial Attached SCSI controller: Broadcom / LSI SAS2116 PCI-Express Fusion-MPT SAS-2 [Meteor] (rev 02)

1 thing I forgot to add. mount_unionfs will also protect your cloud storage since any change will be downloaded first and then any change is done (including delete - which unionfs just hides the file).

Script = bash script. It's the Linux equivalent of Windows / DOS batch file. The "virtual switch" is the GUI. I control my VMs via the GUI.

1. You are thinking from the barebone perspective which can be very different from the VM perspective. So MacOS works with the IGPU, that's great. However, you still have to be able to pass through the IGPU to the VM (in the same way one would passthrough a dedicated GPU via PCIe) + "declare" to the MacOS VM that it is what it is. Those 2 things are usually fraught with difficulties. It could be done with some older gen Intel but I don't remember seeing any success stories with the latest gens. And it's not an Unraid thing. It's a VM thing i.e. any sort of virtualisation envi will have a hard time. 3. Not exactly "hard". You just need 4 VM templates (VM A+, A-, B+, B-, with + having config for the swap disk and - not having it) + the UserScript plugin. Then set up a script that shutdowns A+, starts A-, shutdowns B-, starts B+ (and another one vice versa). I do that all the time with my various VMs. 4. I just don't trust Windows "RAID" aka storage space implementation due to terrible past experience. You may want to rethink your storage layout with focus on using Unraid first, instead of using Windows first. I used to have a Windows-based arrangement too but have gradually moved to Unraid-based which I have found to be superior after the initial time investment in reading up on Unraid (and later watching SpaceInvaderOne videos <-- seriously, new users are spoiled, I used to have to go through pages of the Wiki instead of having someone almost hand-holding through the steps like the vids). 7. As long as it is not running 5.2 GHz all the time then I reckon it's less of an issue. It's almost like tweaking Precision Boost for AMD I guess.

No. It's in the same IOMMU group as a SATA device. My hunch is that it's for the M.2 slot if run in SATA mode so theoretically I can vfio stub it but I have never come to do it since I don't need onboard audio. Of course, ACS Override should work but then in my case, it lags if ACS Override is on.

Not too sure what you are talking about so perhaps some screenshots may help.

The very beginning? As in the login screen? Unless you use non-English characters in your password, I don't see how the keyboard layout would make a diff.

In case you are still confused about the unit. MB/s = megabyte/s = 1 million bytes/s MiB/s = Mebibyte/s = 1,048,576 bytes/s 56MiB/s can be a reasonable speed if your data is not too sequential in nature. It would be certainly too slow if you data is highly sequential e.g. a large ISO image. So it depends. Note though that since you are transferring over gigabit network, your max speed will be limited to 1Gb/s = 125MB/s (8 bits = 1 byte). So you will never get anywhere close to 350MB/s sequential unless there are multiple simultaneous independently connected clients.

1. Can I have the windows VM use the RTX card with hardware access, and have the Mac OS VM use the iGPU in the 9900KS? Highly unlikely. Passing through iGPU is already an iffy affair to begin with. Making it work with MacOS is... I'm tempted to say impossible. You are better off getting a dedicated GPU for the MacOS VM (you might want to google separately whether it would work with a Hackintosh and whether it would work with an Unraid VM). 2. My MB has both USB 2 and 3.0-2, if they are separate controllers I have read that the USB 3.0-2 controller can be passed directly to a VM (in this case the windows VM) For USB hot plug support, and I would like to use a USB add in card for the Mac OS VM will this allow for hot plug on both VMs? Most now-ish chipsets have USB 2.0 and 3.1 on the same controller that is virtually impossible to be passed through. Usually the 3.0 controller is in its own IOMMU group so can be passed through. Note the "usually" since different motherboard, chipset and BIOS version can have very different IOMMU grouping so it's hard to tell for sure. With regards to USB PCIe card, you might have to search on the forum. I vaguely remember a few of them don't like being passed through for whatever reasons. You might also want to consider what I call "warm plug" for Unraid. There is an Unraid plugin that allows you to "plug" USB devices on non-passed-through controller to any VM via the GUI. So you are not hot-plugging but you also don't need to turn off the VM to plug (and replug). Hence my coiled term "warm plug". 3. How would my storage config work? I would like to have the SSDs be VM specific as in one for one VM and one for the other VM, and used for no other purpose other than there requisite OS, I would also like to have the 1TB and 2TB SSHDs be accessible from both VMs but as separate drives, (one to be like a share of documents that both VMs can see, the other a scratch disk for things like video editing and encoding) would this work as specified? 1xNVMe per VM would only work if they are in separate IOMMU groups (assuming you want best possible performance, which requires PCIe passthrough, which needs cooperating IOMMU groups). Often if they are not, ACS Override should help but again, no guarantee. Share doc is easy, just unassigned device mount and smb share. Scratch disk is not easy. To share the disk, it has to be done via network share, which is usually not good for scratch disk due to added latency. If you ata-device-pass-through then it can only be used by 1 VM at a time. 4. How does intel optain work with unraid? I would like to have both of the 8TB HDDs (In raid 0) and the optain drive forwarded straight to the windows VM, this would be a mass game storage area. Don't do RAID-0. Mixing RAID with Unraid is asking for troubles. Optain won't work in a VM the same way it would with barebone. You might as well consider it like a super-low-latency NVMe SSD. 5. I will have RGB ram in the system that I would like to control from the windows VM, is this possible? (Corsair iCue) (all MB RGB will be disabled) Don't know. If connected through USB then probably yes. Any other way = highly unlikely. 6. I plan to use things like the Corsair commander pro for fan and RGB lighting control via a USB 2 port, I understand that I can forward specific USB devices to VMs even if the controller is assigned to unraid is this correct? (All MB RGB will be disabled) Probably yes. 7. How many cores/threads should be allocated to unraid in this case? Don’t plan to have anything other that the two VMs running witch will have there own cores, I’m thinking 5C-10T to the windows VM, 2C-4T for the Mac OS VM, and 1C-2T for unraid, Note that I plan to have an all core 5+GHz overclock present, are there any problems with this config? Your plan is fine, except for the overclocking. You should not be overclocking Unraid (beyond the automated Turbo Boost, which Unraid has a plugin for you to control). Stability on a virtualised environment trumps any speed gain you can get beyond Turbo Boost. Particularly, RAID-0 is already bad enough. RAID-0 + Overclock = you might as well plan to lose data. Last but not least, check out this topic and watch SpaceInvaderOne other guides on Unraid before committing.

这里没人会中文 😅

1. Should be good enough but just barely. 800 single-thread passmark = you might run into the rare case of your CPU bottle-necking transfer speed under the right (wrong?) circumstances. 2. It depends. For direct play (i.e. no transcoding) it should be sufficient but note (1) above. If you need transcoding (i.e. your media format isn't supported by the Shield) then forget it unless you are still in the DVD era.

Where are you doing the copying from and to? In other words, Where are the "FreeBSD...iso" and "highsierra.img" saved? Where are "Downloads" and "Linux" located? You screenshot looks to be MacOS-based. Are you copying over wifi / wired network / br0 bridge? Are you able to test on a Windows-based machine?

No need to remove / reinstall. If you don't touch them, they do no harm.

Not sure where you read about the U9 not suitable for WX chips. Based on the reviews I read (e.g. below), it's entirely sufficient i.e. it didn't thermal throttle with 2990WX load. https://www.phoronix.com/scan.php?page=article&item=amd-2990wx-cooling&num=5 You won't be able to overclock (which perhaps was how the "not suitable for WX" conclusion came about) but then you shouldn't be overclocking Unraid. About water cooling: the only water-cooling you can trust is one that you build yourself. Otherwise, avoiding hoohahs like the Enermax gunk problem is more or less hope-and-pray. And even without a major issue, AIO performance will degrade more over time (compared to air cooling) due to unavoidable liquid lost.

Need clarification: so is the RX 580 primary (i.e. what UNRAID boots with) or not? They are two vastly different situations.

Exactly. I always recommend people to try things out with their existing hardware first before thinking something else is better. For example, it's pretty easy for a Youtuber to pixel-peep and make a fair assessment that the quality is terrible. Yet, when viewed in totality on a 40" TV at a typical viewing distance, very few people have enough visual resolution to pixel peep.

I think you completely missed the latter part of my paragraph. I recommended booting in Legacy mode NOT because Unraid can or cannot boot flawlessly in UEFI with his mother motherboard. Rather, booting Unraid in UEFI has been known in some cases to cause some Nvidia cards in the primary slot to detect that it has been passed through to a VM and thus leads to the dreaded error 43. There are certainly benefits of booting in UEFI; however, for the vast majority of Unraid users, those benefits vary from "can't tell the diff" to "negligible". Yet, the one error code 43, if it happens, would cause plenty of grief - hence my recommendation.

Yeah, that looks better. While you are waiting for the card, I recommend doing these: Boot Unraid in legacy mode. This has been proven to help with the dreaded error code 43 with Nvidia cards. Check your BIOS to see if you can change which slot is used as primary GPU (i.e. what Unraid boots with).

You said "Waiting for the new graphics card (GTX 1660 Super), for the first slot, and GT 710 to the third slot." So is the 710 currently in the 1st or 3rd slot? Without the 1660, you will have to post the IOMMU when the GT710 is in the 1st AND when it's in the 3rd slot (so we can see what groups those slots are in). Just based on your current post, all I can say is you won't be able to pass through whatever slot the 710 is currently in because it's in the same group as the LSI card. (see below) IOMMU group 2: [1022:1452] 00:03.0 Host bridge: Advanced Micro Devices, Inc. [AMD] Family 17h (Models 00h-1fh) PCIe Dummy Host Bridge [1022:1453] 00:03.1 PCI bridge: Advanced Micro Devices, Inc. [AMD] Family 17h (Models 00h-0fh) PCIe GPP Bridge [1022:1453] 00:03.2 PCI bridge: Advanced Micro Devices, Inc. [AMD] Family 17h (Models 00h-0fh) PCIe GPP Bridge [10de:128b] 09:00.0 VGA compatible controller: NVIDIA Corporation GK208B [GeForce GT 710] (rev a1) [10de:0e0f] 09:00.1 Audio device: NVIDIA Corporation GK208 HDMI/DP Audio Controller (rev a1) [1000:0064] 0a:00.0 Serial Attached SCSI controller: Broadcom / LSI SAS2116 PCI-Express Fusion-MPT SAS-2 [Meteor] (rev 02) Of course that is assuming you are not using ACS Override.

Their website said standard doesn't support encryption e.g. LUKS.

In the USB stick, under the config folder, you will have files like Pro.key and Plus.key etc. It should be obvious which one should be the right one since you should know what license you are on. Note: I do NOT recommend deleting the key files. Just rename it to like "Plus.key.old" or something like that would be sufficient. That would help in case you delete the wrong file for some reasons.

A format with the disk in the array will always change parity (there's no "somehow" about it, that's how it should work). So you have lost the data (the formatted disk) as well as the mean to recover it (parity is updated). Now since you stopped it quickly, there is still hope of recovering some data using consumer-grade data recovery software. However, you have jumped on the encryption bandwagon, which is beyond the capability of all consumer-grade software. That means you will have to opt for commercial-grade solutions which are rather expensive. For example, UFS Explorer (what johnnie said) "Pro" license costs €600. And it's not at all a guarantee that you will be able to recover any data. Professional data recovery service gives you the best chance of getting any thing back. However, they tend to cost per GB and I'm certain there will be significant surcharge for encrypted storage. Sorry for being the grim reaper but I think it's more important you know the full picture.

What would prevent you from just try it out and see if you are happy with it? Different people have different tolerant levels to "bad".