thisguyhere

Not sure why this stopped working, but I can't access any IPs on the server's LAN. So I have a number of VMs on the same LAN as my Unraid server (192.168.1.0) but can't ping ANYTHING on the LAN except the server. Unraid server is 192.168.1.10, ping to it works fine. But nothing else on the subnet. Here's my wireguard config. Any help would be appreciated, thanks!

This issue was brought up here, but just kind of ended with no real solution: Using both avidemux or handbrake, I can't get docker containers to utilize the CPU 100%. In handbrake, changed encoder settings from "fast" to "medium" and "high quality" but nothing. Tried with both pinning 4 core+ht and no pinning, but still can't get it to utilize more than a bit over 50%: Even tried prime95 but seeing similar results where CPU is not being utilized fully. This is my third unraid build, first with a E5-1231v3 and then a Ryzen 9 x3900 - which was then subsequently downgraded to the current build (E5-1278v3 + chinese x99 mb). All previous builds cpu would be at a 100% when using handbrake or avidemux. I feel like it's a bios settings somewhere I'm not aware of, or a system wide setting that's missing. Attached diags, any help would be appreciated. jfs-diagnostics-20201110-2212.zip

I would...but I guess just complaining about it made a difference because it's back to normal now. VMs page loads right up with no hesitation. Strange how that works...

2020-07-08 12-05-02.mkv My VM page is slow to load. Is there logging I can enable, or existing logs, to see if I can find anything that's hanging it up? https://player.vimeo.com/video/436556426

everything was bunched up and nvme vm passthrough would not work without it enabled.

Hm...all I did was swap the gpus in each other's pcie slots and it's resolved this issue. Not sure why that would have an impact, but...there you go.

Ryzen 3900x Gigabyte B550 Aorus Pro Hope this isn't too messy. Actually I just tried running Plex with HW transcoding while the VM is shut down, and it definitely isn't working normally. On the previous server, using the same GPU (1050), streams with HW transcode would play almost immediately. Right now, it hangs for a while. And then if I play another stream, nvidia-smi reports only one process despite two streams. Previously all streams would be reported in nvidia-smi. There's definitely something funky going on, will need more troubleshooting.

Tried searching for this but didn't even know how to phrase it right. So this is what's happening. Got a VM with GPU passed in - VM is a nvme passthrough. << This part's working, cool. Got a second GPU intended to handle Plex transcodes. NVIDIA plugin installed, drivers installed, GPU passed into Plex container, Plex setting to enable HW transcoding done. Plex is running, cool. Now, when I play something through Plex and force a transcode, the VM freaks out and start freezing and just being generally unusable. In fact, without even playing anything on Plex, if I run nvidia-smi through terminal, the VM also starts to lock up and freaks out. So it appears the two GPUs in my system are linked and putting one under load is affecting the other? Attached diagnostics. jfs-diagnostics-20200629-1643.zip

Yup, as expected, it was the NIC. Got a intel nic today, everything's working fine now. Damn realtek...

that's what i have my lan ip range set to, just because it's easier to type (ultra laziness). got it, yea the previous (current) server has a supermicro mb + intel nic and it's been troublefree. i'll get a hold of a pcie intel nic and try again. thanks for linking to the other post describing it.

So built a new machine to replace an aging unraid server. On first boot to unraid, IP address is not getting assigned. I know the 169.254.* ip is a result of this machine not being assigned an IP address via DHCP. Some googling yields a number of similar occurrences, but no obvious answer. Things I've tried so far: - Using 3 different types of usb sticks - Unraid versions 6.8.2 and 6.8.3 - network.cfg = USE_DHCP = yes and no - network.cfg = BONDING = yes and no The the one and only NIC is onboard Realtek: I'm preparing to purchase a pcie nic card to see if it's an issue with the onboard nic. But running windows from this same machine, there are NO networking issues (I'm planning to passthrough a win10 nvme to vm). Diagnostics are here: tower-diagnostics-20200623-1050.zip Any insights would be appreciated.

Had to share ... shares going missing issue for me has disappeared completely with the update. Thank you kindly, keep up the great work everyone!

Caaan doooo! Didn't realize 6.6 is stable ... sorry. Will try it and report back, thanks!

good point, i was just waiting for the stable release of the next version - should i just go ahead with 6.7-rc?

I've already read through many topics about shares going missing, but none of it seemed to apply to what I'm experiencing. So opening this new topic - if it's been covered elsewhere, a hint would be appreciated. I'm on 6.4.0, occasionally all of my smb shares go missing. Not found in the UI, not found on the network. But it happens infrequently, anywhere between a couple weeks to a month - roughly. A reboot restores everything, and back to normal life for a while. What should I be looking for in the diagnostics file that may point to what's happening? Do I grab the diagnostic file when the shares disappear? Or after reboot? Both? I've just learned to deal with it because it happens so infrequently. It would be nice to find a permanent fix though. This is the only problem to an otherwise rock-solid server.