IG82

Members
  • Posts

    76
  • Joined

  • Last visited

Everything posted by IG82

  1. Definitely something wrong, adding entries to securetty is only a quick fudge as the numbers increment. No issue with 0 byte/corrupted ssh keys either from what I can see.
  2. I would suggest trying two options which I used successfully: 1. See if you can pass through the motherboard sound device. I did this and used the optical output from my motherboard. 2. Pass through a usb soundcard at device level, not the controller. I use the Libvirt USB plugin for this as it makes it easy to switch between machines.
  3. Try using this https://lime-technology.com/forum/index.php?topic=47240.0 plugin. Identifying and passing through identical devices in the XML is tough unless you can isolate USB controllers. This plugin lets you do it on the fly, I haven't tried identical devices yet but I use it to move things around VMs dynamically and it works great.
  4. Nice one, glad you got it working in the end, I know how frustrating these issues can be!
  5. Marking this thread as solved is pretty unhelpful, solved should only be used when a solution has been found and provided. FWIW I am running a Windows 10 VM with a Gigabyte G1 1070 and haven't had these issues. I am currently running driver 376.33 which is the latest, this has been updated using the Nvidia utility multiple times up to this version with no problems. If you could provide some further details you may get some more help (I am not an expert but happy to try). The use of the ROM file suggests that you are running a single GPU system? I am on X99 but have an additional basic GPU so haven't messed around with trying a single GPU yet.
  6. Try setting the USB controller to USB3/XHCI. If you can't do through the wizard I think the model in the XML would be nec-xhci instead of ich9-ehci / ich9-uhci. I think I have had similar issues with USB devices not being available at boot, either this helped me or it was the other way around and setting to USB2 which you already have.
  7. I vaguely remember coming across this and getting round it, unfortunately I didn't document it so it may have been another problem. Take a look at this thread, which I think covers it. It has details of some xml changes, but from the shell try: fs0: cd boot/efi bootx64.efi This question would probably get better support in the 'VM Engine (KVM)' part of the forums, but hopefully the thread helps!
  8. Thanks, was inclined to think that too, but was hoping there was a reason not to worry and save replacing it!
  9. Wondering what the consensus is here, I have 2 end-to-end errors showing on an out of warranty Seagate 4TB NAS drive. They occurred while I was copying data off to another disk, so I could convert from reiserfs to xfs. No errors were reported within the MC session during the copy The rest of the data (another 1.5TB) copied with no further errors to the drive An extended SMART test completed with no further errors I converted to xfs and ran a parity check with no further errors Just wondering if people would straight off replace the disk, or monitor it for further errors? Thanks Ian
  10. I recently decided the same with 2 x N36Ls, that said you can definitely fit 6 drives in them. One of them had a 2 port adaptec SATA card and the other I just used the ODD and eSATA ports. With a cache drive plus 6 hard drives you will need an expansion card I think. Longer term I would go for something larger, but in the meantime something to consider!
  11. Just experienced the same with the beta and very little by way of a report after the 5 runs. Think I will stick with the regular script from Joe and screen without messing around with plugins.
  12. It is not completely clear what you are looking for. Do you have hardware you want to use for the purpose or are you looking at what you would need? I would say it is possible, although to be completely honest I haven't done what I call extensive testing yet. My long term goal is to have unRAID with two graphics cards at least, one for my primary gaming PC connected directly to monitors and also streaming. The second card would then be used for my girlfriend to stream when she wanted to play something. I am pretty convinced it will work, I got an HTC Vive up and running and that was better on the VM than my baremetal machine. Admittedly the spec was better on the VM (780ti vs 7970CF) but certainly no issues with input latency on the motion controllers. I also have a Steam Link which I hope to use a bit today as I have a day off. Will also be testing PC streaming to 1440p/4K monitors at some point too. Get the right hardware for unRAID KVM, plus something like a passive Intel NUC/Gigabyte Brix and you should be able to do what you want. This would be using your existing network cabling rather than messing with video over IP.
  13. Honestly not as much as I would like. Most of the content is a bit too much like tech demos. That said i haven't noticed any control/input issues compared to metal. I need to get back into Elite and have a proper play session, but that obviously wouldn't use the motion controllers as input. It didn't work when passing through the USB devices individually in the GUI, of which there are about 10! But when passing the full controller all was good. With my motherboard (Gigabyte X99 SLI), I had to tweak bios settings to get separate controllers which resulted in them being USB 2 only. I haven't tried USB 3 yet with any passthrough so this may be a consideration for the Rift which I believe requires it. I was getting driver reset issues which causes the headset image to freeze. This required resetting the headset/software, although the VM was still fine. The main cause here was I hadn't enabled MSI on the GPU/audio in my eagerness to try it out. Also I backed off my overclock a bit on the GPU and haven't had any issues since. All in all though it has encouraged me to get rid of my on Z68 based desktop and have my gaming pc virtual. As it is under water though it means draining the loop and setting back up which I will struggle to find time to do!
  14. https://lime-technology.com/forum/index.php?topic=48478.0 Vive works
  15. I have had exactly the same issues before, certainly on a VM and I am pretty sure baremetal also. I setup a W7 baremetal machine last night and got bored of waiting a few hours for updates to be found. I had success with the following tip from 'oldsh_t' on the mydigitallife forums: On a clean install I: >manually install IE11 and the updates needed for it >manually install KB3102810 (which is to speed up windows update) >then install 50 updates at a time from the list offered to me by M$ using windows update. I didn't actually bother with the 3rd step, manual IE11 install and the KB update detected all the expected updates. All installed successfully after that. I am making a note to try this next time I install a W7 VM. Hopefully it may help you. Cheers Ian
  16. Works Just got my HTC Vive working under Windows 10 KVM using 6.2 b21. Other than a few small audio glitches it initially seems good. I was running baremetal on a 2500k @ 4.8Ghz with 16GB RAM and 2x7970s, but I think the GPU(s) were not quite up to the task. I have been having increasing issues with Crossfire recently in modern games so was planning on changing. The unRAID machine is a 5820k (stock currently) with 32GB RAM and a 780ti. The W10 VM has 8 cores and 16GB RAM assigned. With the GPU overclocked to 1150Mhz it gives an average fidelity of 7 and ready for VR result. Need to see if some of the odd issues I was having before with GPU passthrough are resolved, but so far looks like 6.2 is a great step forward.
  17. Not seeing PC Mark 7 or 8 give results like you have, are you sure that is the software you used? Is there somewhere specific in the results to get the CPU / GPU breakdown, I may be missing something. Happy to try and run comparisons if you can give me specifics, I would guess you have something wrong with your setup/installation on the VM.
  18. I have been doing the same recently and haven't found such a big difference although I was using the Asus Realbench suite. The GPU based benchmark was the same, the CPU was about 95% if I remember. Let me try setting up the same machine again and run the same as you. My testing was on an Intel X58 platform, will also test on a new X99 setup as I should be able to tweak the xml. Afraid I don't have anything AMD (other than an AM2 setup).
  19. Given the consensus seems to be that 2GB with nerdtools is insufficient I stopped the memtest runs. Partly triggered by the old TV I was using as a monitor giving up the ghost too! I don't have anything else with VGA easily to hand, need to get a cheap monitor for my servers. It had run 8 or so completed passes with no failures so should hopefully be ok. Have removed nerdtools and am running similar transfers over again.
  20. Hmm, the requirements are still listed as 1GB. As per my opening post I have installed the nerdtools plugin but otherwise no plugins, docker is not running, and no virtual machines. 1st pass of memtest is clear with no errors but will leave it running until tomorrow after work so it gets a good few runs in. If the memtest comes up clear I can try bumping the memory as my workhorse server has 60GB ECC so I can spare a 4GB DIMM I think! Would be great if someone could cast their eye over the syslog, pulled out of the zip and attached separately this time. syslog.txt
  21. Thanks guys, very familiar with running memtest and the method, just for some reason it did not occur to me that this was a good first step this time!
  22. On one of my N36L microservers I have been getting out of memory errors. It is running a modified bios to enable full speed SATA on the external/onboard ports, with 6 x WD20EARS/X drives and 2GB of RAM. Brand new installation of 6.1.7. Only plugin installed is the nerdtools pack. Not using docker or any VMs. I am copying data across to it via by mounting and writing to an NFS share via MC. I can see in the syslog it had errors, although this time around didn't experience any issues. Last time though the webgui had hung and I could not access any mounts/shares. Although I could telnet into the server, I was unable to do anything under /mnt. I ended up rebooting it which resulted in a clean shutdown, hence the parity check in the new syslog. Any ideas or suggestions on what the cause is? I haven't noticed it happen on my other N36L server which is currently running 6.1.6, but I haven't done the same kind of large copying of data to that one for a while. Attached is a diagnostic zip. Thanks Ian tower-diagnostics-20160123-2358.zip
  23. Woohoo! Yes Tom, that resolved the issue. Many thanks for solving the issue. Cheers Ian
  24. I noticed that too, have never had a cache drive in this machine. The drive in the vars file may have been in the server but it would have been a long time ago! I would have moved from 1TB WD to 2TB WD then 4TB Seagate. I have tried changing the cache drive to no slots on the main page but this makes no difference. It's very frustrating, 5.0.6 had a persistent issue with not shutting down cleanly. No plugins at all as I run a separate XenServer, no solutions in the previous support thread. Now I cannot perform an upgrade to the latest version.
  25. Hi Tom, Full copy/paste is too long (>20K characters) so I have attached a txt file with the vars content. Thanks Ian vars.txt