• Posts

  • Joined

  • Last visited

Everything posted by jsebright

  1. Great work guys. When can we see an RC for 6.11 ? 😁
  2. Did a scan of the flash and it came up OK, so am saving that option for later. I swapped cards that I need (10gb & drive controller), and removed a PCIE usb card that I can swap for another in a bit. I've finally got the "Interface Rules" back and can assign eth0 and eth1. I've not yet been successful in splitting the IOMMU groupings yet and have been trying unsuccessfully to get that to work. This is probably where some of the issues came in. I have also experienced problems with network not coming back up after waking from sleep - probably related, see this post: Going off topic. My ideal setup is where I can use onboard (intel HD930) graphics for a VM and perhaps some transcoding (quite rare), pass through a usb hub to the VM, Fibre connection as main, with gb for waking from sleep, server will go to sleep at night and wake up OK when needed. Stopping that is that I can't manage to split the audio part of the HD930 out into it's own IOMMU (so am using an extra card), and have been having problems with the network cards not working - seemingly when I've got the USB controller in (that was being stubbed). Possible issue remains is that some combination of hardware has prevented the Network-Rules from being created / maintainable. I'm going to try a different USB card and could try more moving around of cards to see if I manage to break the audio out and don't need the additional graphics - although I'm sure that will cause problems with getting VMs to use the onboard.
  3. Time to re-create the flash drive? I'm not on the "My Servers" yet - will that slow down provisioning a new one if necessary. Appreciate the swift help. Thanks.
  4. renamed networked.cfg - I can't see that a replacement has been created, but I haven't re-configured anything. No network-rules.cfg either. I'd been looking for a config file thinking there might be an issue with it but couldn't find one. I've recently swapped motherboard (and cpu), so the mac address of eth0 will have changed.
  5. I've made quite a few changes recently and have been up against this issue for a while - I have two network adapters and can't point the right one at eth0. Onboard is gigabit, and I have a 10gb fibre card that I want to be the default. I want to have the onboard activated so that I can use wake on lan which doesn't work with the fibre. I've managed to lose the option to assign eth0 and eth1, and can't get it back. Have tried reverting to rc7 and back, and many many reboots but it fails to show. It's left me with the wrong adapter as default. Diagnostics attached. Goodness knows what horrors lurk within.... tower-diagnostics-20220512-1456.zip
  6. Oh yes. Very similar issues. Thought it was just me and continuinually fiddling with hardware that broke it. I was finding the server wakes up from a wake packet, but I couldn't get it to work. Having a keyboard attached and manually logging in and doing a reboot, gave a clean reboot. But I had to do that blind as my screen freezes on the login prompt (even before sleep) - presumably due to having intel_gpu_top plugin installed. I've set it to shutdown instead of sleep - takes longer to "wake up" but is more reliable. Would be nice if we could force the networking to restart on waking.
  7. Same here - not sure if it happened in rc3 or rc4 though. Changing ownership of the appdata files (just for misbehaving dockers) to nobody:users seemed to fix it for me.
  8. That's correct, I did. Struggling to get it to work again though. Going back to VNC, uninstalling all drivers & old devices before adding the graphics will probably help.
  9. I haven't - that's where it all goes wrong. You're stuck with the default drivers. I'm on 6.9.2 and not ready to try 6.10 yet - still waiting for the next RC, "soon?". I'm hoping that updates to KVM/QEMU will help, as may future bios updates, and updated drivers from AMD. For now I've reverted back to adding a physical graphics card for VM work. A bit disappointing.
  10. Sorry, no idea where the xvga='yes' came from. I try to stick to the Forms mode when editing VMs. As I said, for me I got to a state where there was no display but the VM was running and able to serve a remote desktop. I also saw the Microsoft Basic Adapter, but just did a reboot and the screen came up. I'm not 100% sure if the output is from HDMI or DP>HDMI adapter (I'm then running two long HDMI leads to another room). I don't get multi monitor from the basic adapter so it might be that only one of the outputs is working. If you are able to test a display port connection it might show something. I'm not seeing any errors in the logs. Sorry I can't offer more help - I don't understand enough to explain what's going on.
  11. <hostdev mode='subsystem' type='pci' managed='yes' xvga='yes'> <driver name='vfio'/> <source> <address domain='0x0000' bus='0x0c' slot='0x00' function='0x0'/> </source> <alias name='hostdev0'/> <address type='pci' domain='0x0000' bus='0x04' slot='0x00' function='0x0' multifunction='on'/> </hostdev> <hostdev mode='subsystem' type='pci' managed='yes'> <driver name='vfio'/> <source> <address domain='0x0000' bus='0x0c' slot='0x00' function='0x1'/> </source> <alias name='hostdev1'/> <address type='pci' domain='0x0000' bus='0x04' slot='0x00' function='0x1'/> </hostdev> <hostdev mode='subsystem' type='usb' managed='no'> <source> <address bus='1' device='6'/> </source> <alias name='hostdev2'/> <address type='usb' bus='0' port='1'/> </hostdev> This is the section with the graphics - all the "hostdev" bits. The bit I edited was the line <address type='pci' domain='0x0000' bus='0x04' slot='0x00' function='0x1'/> I don't have the graphics & audio bound to VFIO. I don't think that the whole script will help - I've not got a full understanding of what it does, and it's probably a bit of a mess due to previous changes to this machine. The important points are Q35 Seabios, no other devices (add them later), get remote desktop working first (whilst graphics are VNC) so you can log in remotely if the display doesn't work (after you've changed the graphics) then you can check device manager and reboot.
  12. OK, so I couldn't resist having another go. This is what I've done and the issues Made sure the AMD Cezanne graphics and driver were uninstalled from the VM Added the GPU and audio to the VM Fiddled with the Advanced settings xml so that the passed through audio bus was the same as the graphics, and the function was 0x1 (sort of matching the devices pre-passthrough). This might not be necessary if we're not installing drivers. Booted up the VM. Black screens and nothing. But, I could remote desktop into the VM, so I did that and rebooted it. It came up with a screen. Made sure the keyboard & mouse was attached, and logged in. Rebooted - came up straight away with the screen. Changed the resolution to match the monitor, rebooted and it's still good. Shut it down, then rebooted and all fine. I even tried pausing the VM from unraid then starting it and that worked. So - quite successfull. But - I can only see one monitor and I have two plugged in (I think both connections are plugged in at the server anyway!). Display adapter in Device Manager shows "Microsoft Basic Display Adapter". Obviously not making the most of the chip, but it's perfectly usable. I'm not going to go further and try to install any driver updates as it's got a high probability of jamming something and the server might require a hard reboot. It's possible to get this far. Good luck!
  13. Can't remember exactly, but VM is Q35-5.1 (latest Q35 for 6.9.2) with SeaBIOS. I think that some time ago I had to move from i440fx to Q35 to get a GPU passthrough working. I've not tried much recently as I don't want to crash the machine again, but I think the first try I hadn't even done the VFIO binding. Tempted to have another try to see if I can get something basic working without the drivers that doesn't crash the system following a VM shutdown / startup. Will post back when I do that.
  14. Have just upgraded to this CPU with the aim of removing a graphics card and reducing power consumption, but still hoping to pass through to a Windows VM. Initial passthrough worked OK (I think). (I have not tried to mess with the GPU bios.) Then trying to install the drivers I had some problems. Driver install seemed to hang, then once I'd got round that (possibly), then I found that turning off the VM caused horrible problems with CPU usage going crazy and the system becoming unresponsive. A reboot was the only solution. Since then I've avoided this hoping that there will be some fixes in a new driver, or in 6.10 (not ready to try a beta build yet as I only have the one server). I have meddled with an Ubuntu VM, but couldn't get that to pass through. Am tempted to try doing a clean windows install to see what happens, but am wary of locking everything up again, also might try and see how I can get on without drivers as I'm not worried about gaming - just occasional desktop use. I hadn't considered dumping the bios might help as I haven't needed to go that route before with a stand alone (very old) AMD card that I passed through. Don't really want to put the old card back in, but that thought has crossed my mind. Interested to follow this and see what progress is made.
  15. ing Chia... ey at path: /root/.chia/mnemonic.txt ey at path: /root/.chia/mnemonic2.txt ot directory "/plots". ot directory "/plots1". ot directory "/plots2". ot directory "/plots3". ot started yet daemon vester: started mer: started l_node: started let: started annot create directory '/root/.chia/flax': File exists ing Flax... ey at path: /root/.chia/mnemonic.txt ey at path: /root/.chia/mnemonic2.txt ot directory "/plots". ot directory "/plots1". ot directory "/plots2". ot directory "/plots3". ot started yet daemon vester: started mer: started l_node: started let: started ing Plotman... ing Chiadog... Chiadog... ing Flaxdog... Flaxdog... Machinaris API server... Machinaris Web server... d startup. Browse to port 8926. I might have restarted the docker since - not sure sorry. With the usual first few characters cut off - If you know where I can get all the output I'll update it. - still getting plotting failures. This page https://github.com/madMAx43v3r/chia-plotter/issues/574 seems to suggest it may be RAM related. I have had some other crashes with the intense processing (otherwise the server is stable, but only lightly used), so I might do a full reboot and see what I can tweak.
  16. Sorry, that's something that took me a while to get right, and I couldn't see what was wrong with it. I'd suggest you try to revert to defaults and as simple as possible. Check the drive paths are mapping to where they should be. Can't offer any more help as I don't know enough...
  17. @localh0rst I had the internal server error. Logged into the docker and ran "flax init" that seemed to start it for me. May or may not work for you...
  18. Ah yes. Why didn't I think of that? Has the advantage that it doesn't get killed if the docker restarts. Watching a few together in just one window is pretty good though, and a bit less to keep track of.
  19. Nice. Haven't seen that before. I couldn't seem to open two windows of the unraid docker console, but I did find I could just use the command watch -n10 du -sh /plotting /plotting2 and it watches both folders in one window.
  20. Thanks for your reply. Seems like the runs are crashing sometimes. The tempdirs have files from different runs leftover in them. I've just checked my config against the wiki sample (was doing that as I saw your previous post). I've got max jobs set to 1, but had the staggers set differently - might have been causing the issue. I've got the threads set to 4 - I've only got 5 pairs of cores available (Ryzen 2600 with a pair pinned in case a VM wants a look-in) and it does tend to run them at about 90%. Perhaps this needs taking down from the default? - I'll see how it goes with the default stagger options before reducing it. Thanks again for everything your doing with this.
  21. It also took me some time to get MadMax plotting to work. I also think there's an issue that it's not clearing temp folders up. I've got two ssds that I'm using and have one of them as the primary temp, and one as tmp2. They are slowly filling up until plotting stops. After it's all ground to a halt - see plotting and plotting2 : root@Tower:/# df -h Filesystem Size Used Avail Use% Mounted on /dev/nvme0n1p1 932G 673G 259G 73% / tmpfs 64M 0 64M 0% /dev tmpfs 32G 0 32G 0% /sys/fs/cgroup shm 64M 60K 64M 1% /dev/shm shfs 3.7T 893G 2.8T 24% /plots /dev/sde1 447G 313G 135G 70% /plotting /dev/nvme0n1p1 932G 673G 259G 73% /id_rsa /dev/sdf1 447G 447G 24K 100% /plotting2 /dev/sdb1 7.3T 7.2T 153G 98% /plots2 /dev/sdc1 7.3T 7.2T 153G 98% /plots3 tmpfs 32G 0 32G 0% /proc/acpi tmpfs 32G 0 32G 0% /sys/firmware I'm having to stop plotting, clear the files, and start it off again. Don't know if this is a MadMax or Machinaris issue. @guy.davis ?? This is still faster than the chia plotter, I don't have much diskspace left...
  22. Just did a "check for updates" and it's available. Just waiting for some plots to finish. Looking forward to the latest version.
  23. What's the best way to upgrade the unraid docker if it doesn't show "upgrade available". Will an edit of the settings pull the new images? PS: many thanks to @guy.davis for making and supporting this. Good luck with your farming.
  24. This problem occurred again, then I think I worked out what was going on. The device was "disappearing" when I started a VM, but only a certain one. I had had to fiddle with it a day or so ago as it wouldn't start. Something must have got messed up meaning the VM was trying to take control of the nvme drive. I could spot the device in the xml, but am not confident enough to edit it. Just saving the VM settings from the forms view didn't clear the device, but selecting all the possible usb devices and the one pcie device, saving, then clearing them all and saving seems to have sorted it out. Thanks for the support - I know a bit more about checking disks now.
  25. Am up to date on Bios - a reasonably new one that's been in for a few weeks before this issue. Have added the script. Will fix the other issues and see how it goes. Thanks both.