space

With turbo write enabled I seem to get an extended period of good speed at the beginning, then dropping to 8-9MB/s, and ut to 50MB/S again, like the wave pattern above. Still better than before I guess.

Will check out turbo write! Thanks! (I have SSD cache disk).

Is this all I can hope for? This is copying a large file from my win10 VM to a share on the array. Low spots are at around 5MB/s. Disks are 1TB SATA. Home network is 1Gbit, if it's using the network instead of communicating directly?

Ok, so what I posted nearly a year ago was not quite the truth. It became much much better, but still there were some issues. Yesterday I picked up a new usb mixer / soundcard. The Yamaha MG10XU. This caused the audio problems to come back tenfold. Installed the Yamaha driver, tried every usb port (both controllers), but nothing helped. Audio playback was unusable. Tried a new (for me) way to pass through the USB controller, editing the syslinux configuration with the vfio-pci.ids parameter istead of editing the VM xml-config: append vfio-pci.ids=8086:8d31 isolcpus=4,5,10,11 initrd=/bzroot and then selecting it in the VM-config. No luck. I was close to quitting the whole unraid story, but as a final test I bought a USB 3.1 PCIe controller card today. (This: https://www.komplett.no/product/911585/datautstyr/pc-komponenter/hovedkort/tilbehoer/st-lab-usb31-gen-2-2xtype-a#) Plugged it in, connected the Yamaha, to identify it, and added it to the syslinux configuration along with the onboard controller. append vfio-pci.ids=8086:8d31,1b21:1242 isolcpus=4,5,10,11 initrd=/bzroot Selected it in the VM vonfig and booted up my WM again. PERFECT sound! Even on very low buffer-size in ASIO-settings. What a bliss! Perfect playback, and perfect recording in Cubase. OK, so it's day one, but I really think I have solved my issues this time. This means: If you have problems with corrupted sound of your USB soundcard, an external PCIe USB controller might be a cheap and easy solution. No matter what advice I followed, it didn't completely solve my issues, just make them "livable". (isolating cores, MSI tweaking, xml-tweaking, win10 tweaking) As seen in the first post of this thread, my hardware is: Intel Core i7-5820K, Socket-LGA2011-3 ASUS X99-A, Socket 2011-3 ASUS DC2 OC NVIDIA GeForce GTX 970 4GB Crucial Ballistix Sport 16GB 2400MHz DDR4 DIMM 288-pin Unraid 6.3.5 Latest main board firmware (4302) Latest virtio drivers. (0.1.126)

I had kind of the same issue. My VM was running on the cache drive, and had free space to spare, when looking at it from within the VM, and the VM configuration left space on the cache drive for other cache activities. Or so I thought. Looking closer the cache drive had 0 bytes left. No wonder things were acting up. Moved the VM to a disk share, and all worked well. Then swapped the cache drive for a new larger drive, and copied the VM back. No issues!

Sounds familiar. Internal soundcard lags behind after a minute or two. External usb souncard has no issues (at the moment). Refreshing browser, and jumping the timeline resets the problem (for the next minute). repeat.

Your system is quite similar to mine. (Asus) X99, 5820k, gtx 970 and gt 210 and scarlett 4i2. USB pass through on one bus. No issues at the moment, though I have had massive problems with stuttering audio. Each new firmware update on my X99 seems to handle things better. Have you updated your motherboard firmware? I have never seen that kind of lag you are experiencing though.

Is the task manager in the VM also reporting to much memory? I get these "out of memory" warnings in win10 too, even though task manager reports having 2-3GB left of the 8GB assigned RAM. My box is running 16GB total RAM, and at the moment I'm just running this VM, and plex docker (idling). Tried this?: I have done those changes now. Will see how it works. (Edit: I just did RAM x 2, not x3 as he states. Win 10 was previously set up according to the unraid recommendations in the VM setup guide) 19/05/2017: No issues since the changes above (in march).

Just to follow up, I first did the procedure for removing a cache pool disk. Disconnected that disk and inserted the new larger disk. Then did the replace cache disk procedure, and disconnected the last of the old cachepool discs. Works like a charm.

Thanks! ...And of course. I ment set it up as a cache disk. It's not much data at the moment. Just a small VM and some docker/Appdata files. My main VM is going back on cache (ssd) when the new disk is up.

What if you want to swap out 2 old small ssd's with 1 new larger one? I'm swapping out 2x100GB now, and settled for 1x250GB. Can I just do the folloing: 1. power down 2. disconnect old drive nr.1 3. insert new drive 3. power up 4. set up new disk as cache only 5. power down 6. disconnect old disk nr.2 7. power up 8. done?

Dropbox is making this hard. It has to be installed on the VM, including the file storage. Maybe making a third VM disk is a solution. Like my seperate games-disk.

Thanks! I think I'll go for the bigger cache pool then.

My win10 installation is 60-ish GB, on a 100GB image. Kind of basic install with Adobe Lightroom, Photoshop, and MS Visual studio. I guess that would make my minimum size for a VM 80GB, to leave space for updates etc.

Is it a good idea to add a SSD to the array (as disk 4), and set up a share (for VM's) to only use that disk? Then also specify all the other shares to use disk1, disk2 and disk3 only? (No failover) Or is it better to just swap out my cache disks with 2x250GB SSD, and run the VM's of cache only? (maybe easier to manage?) What is best practice?

Thank you! I have been struggeling with the anniversary update since it was released. I should have searched this forum sooner... One core did the trick!

Thanks! I will look into that.

I start of with a screenshop of my disk setup: So, I had a lot of problems with my Win10 VM. I was running it off the vdisk share that was set to cache only. After lots of frustration I finally found out that the cache disk was running out of space. (I wasn't thinking of the other things that uses the cache disk when assigning VM disc space.) Anyway, moved the image to a share on the array, and it's been working with no bigger issues since then. However it is much slower than when on the cache SSD's. My VM image is 100GB now. I use some large software, and I think it looks like a hassle to put some programs under program files into a seperate image. (My games are on a image on the normal share though.) What would you do in my situation? I'm thinking of making the cache 1 SSD only, and move the image to the other SSD (is that even possible?) My main goal is to have my Win 10 VM on SSD. My installed files are now 60-ish GB, so that means I can shrink it to 80GB as a minimum. Will it then run fine in cache only given my disc setup? What are my options?

It might have been a windows update or something, thar caused the system to try to reboot. I will follow up if it happens again.

For some reason my win10 vm crashed again today, while i was not using it. The log in Unraid stated this: 2016-09-02T19:50:51.137856Z qemu-system-x86_64: vfio: Cannot reset device 0000:00:14.0, no available reset mechanism. ... ... 2016-09-02T19:50:51.137856Z qemu-system-x86_64: vfio: Cannot reset device 0000:00:14.0, no available reset mechanism. 2016-09-03T19:21:18.578179Z qemu-system-x86_64: vfio_err_notifier_handler(0000:02:00.0) Unrecoverable error detected. Please collect any data possible and then kill the guest 14 is the USB-controller that I have passed through. Is this some common error? What can I do to prevent this from happening? Edit: Code from XML: <hostdev mode='subsystem' type='pci' managed='yes'> <driver name='vfio'/> <source> <address domain='0x0000' bus='0x00' slot='0x14' function='0x0'/> </source> <alias name='hostdev3'/> <address type='pci' domain='0x0000' bus='0x00' slot='0x0a' function='0x0'/> </hostdev> Main board is an Asus X99.

After much trial and error everything now seems to be in perfect order. Problem with unraid locking up on win10 VM boot solved by going beta, and now RC-version. Now it runs great, without any latency, pops or cracks in the sound. Tuning the VM XML in every way, never getting it quite there. Solution: Latest Asus X99 bios update. I now even run an unmodified XML. What a bliss! And a big thanks to everyone in this thread, and the forum in general, for advice, tips and tricks!

This is my biggest issue as well, so a solution would be appreciated. I how now upgraded to 6.2 rc3. I hope that helps. Last time Win10 locked up it was displayed with a pause symbol in the web gui. I have disabled automatic updates, so Win10 should not reboot itself. (Previously (on unraid 6.0(?)) it locked up the whole unraid server every time windows update booted the VM...)

Thanks! Lots of things added I see. VM now starts, but USB passthrough is broken, so no mouse or keyboard action. My USB controller was listed under "Other PCI devices" as Intel C610/X99 series chipset USB xHCI Host Controller | USB controller (00:14.0), which is correct. It still is listed in the XML as: <hostdev mode='subsystem' type='pci' managed='yes'> <driver name='vfio'/> <source> <address domain='0x0000' bus='0x00' slot='0x14' function='0x0'/> </source> <address type='pci' domain='0x0000' bus='0x00' slot='0x0b' function='0x0'/> </hostdev> like before. Well, Address type is different. Old code: <address type='pci' domain='0x0000' bus='0x00' slot='0x07' function='0x0'/> But this is auto generated, right? Edit: Solved: Removed all hostdev tags in the xml except usb controller. Started VM. (no screen, but mouse lighted up) Stopped VM. GUI edit VM, and choosing GPU. Save. Opening XML. Voila! New tags <alias name='hostdev1'/> and <alias name='hostdev2'/> Started VM. Everything working again.

Do I have to edit my win10 VM config? Updated from last beta to RC3. Everything works (Also xubuntu VM) except Win10 VM (with gpu passthrough). Execution error internal error: early end of file from monitor, possible problem: 2016-08-02T11:32:03.472214Z qemu-system-x86_64: -device vfio-pci,host=03:00.0,bus=root.1,addr=00.0,multifunction=on,x-vga=on: vfio: error opening /dev/vfio/29: Operation not permitted 2016-08-02T11:32:03.472246Z qemu-system-x86_64: -device vfio-pci,host=03:00.0,bus=root.1,addr=00.0,multifunction=on,x-vga=on: vfio: failed to get group 29 2016-08-02T11:32:03.472254Z qemu-system-x86_64: -device vfio-pci,host=03:00.0,bus=root.1,addr=00.0,multifunction=on,x-vga=on: Device initialization failed

I used this tip to turn off updates: http://www.thewindowsclub.com/turn-off-windows-update-in-windows-10 It gives me a notify instead of booting at free will. This is only a quick and easy fix for now, to see if that is actually the problem, or if the system locks up for some other reason. I guess I will know in about a week. I think i have maximum uptime of 15 days on my server before something crashes it completely. I also stopped plex media server and sabnzbd. (I know. stupid to do three things at the same time...)