Demon1337

I have finally fixed it. I found that the templates from the old USB drive were not there and without them moved over the containers would not boot. I had attempted to load the same containers I had and when I did the first one all the data was deleted in the process. Thankfully it was the test container. For anyone that has this issue in the future: You will need the templates from the containers. They exist at: /boot/config/plugins/dockerMan/templates-user Attached is a version of the template in case you need it to remake yours. You can also get the template from another container loaded fresh. The important part is the network information. In my case br0 and the IP. but you should try to get as much information as you can. my-PlexMediaServer.xml

Version: 6.12.6 Hardware: HP Proliant DL380 Gen9 Network Card: 10G PCIE GPU: Tesla M40 I have recently had issues with my server. The USB stick just stopped booting into UEFI mode. It was an old stick so it having issues is expected. I loaded a fresh OS onto a new USB drive and things were back going. Fresh OS, Fresh USB. 24 hours later, the server just locks up and wont let me reach any of the containers except with ping commands. VMs were the same story. Tried login at the box itself. No dice. Gives an error saying: -bash: /usr/share/bash-completion/bash_completion: input/output error I reserve that I need to reboot it. So I do that but unplugging for a while and coming back to it. Server boots and most things look fine. It was a bad reboot, but things work. I start looking into issues related to 6.12.6 and find something about MACVLAN. I don't see anything enable for me for that, so I just go about installing the typical plugins. 1. Fix common problems 2. GPU driver 3. Nerd Tools Fix Common problems picks up MACVLAN as an error. Ok, I set out digging where. I properly shutdown the array and bridging is enabled, which means MACVLAN should be disabled. I spin back up the array and look at the docker containers individually. One of them is set to bridge instead of BR0 (I IP all of my services) I delete that container as it wasn't needed anyways. Suddenly I look at all my containers are off. Powering on gets an error saying it doesn't exist. Logs say this for every container: Namecheap-DDNS: Error response from daemon: network dd7e9e2d8f64eceeee342e267c8acb612a734b51ffde7059005b70b0251a6e9f not found Listing networks shows: NETWORK ID NAME DRIVER SCOPE 40dd8737aab6 br0 ipvlan local 214e8ee49a31 bridge bridge local 1f1179c5678d host host local a013fcade590 none null local Listing the containers...completely empty. I would think I could just create that network but not sure exactly how that would be done, or how it would respond to the network. No idea if it just picks up my containers. I have no clue exactly what is going on at this point and am at wits end with the whole thing. I have has zero issues out of Unraid since March out of this box. No errors or anything. The old USB stick ran for something like 4 years without issues. I am just frustrated that something obvious seems broken and I can't quite put my finger on what went wrong. If anyone has any clue exactly whats up here, I am all ears. I fear if this keeps up I might have to change what OS I use because this is really starting to become a large pain point. unraid-diagnostics-20240101-1759.zip

changed that and no difference. still only going at 1G

eth0 is a specific interface. It doesn't let you remove it or change which interface it is. At least not from the GUI. That is a Linux setting and eth0 corresponds to which physical interface what that MAC address.

I am almost positive I have figured out the issue. Eth0 is the default connection of the system and it will not let me change/remove that. Eth0 is a 1G port that is always set as the active member. Meaning that no matter what changes it will only ever communicate at 1G. Guess I have to goto the console to try and figure out how to get Eth0 off the network members.

Nothing has changed with the cables until this issue happened. The DAC and fiber are both pretty new. having been in for less than a year or so. If the fiber were an issue it would just not work, and I have had zero issues out of my desktop. But more over, I shouldn't be seeing issues within the networking of the Unraid server. I should be able to get the full 10G speeds if I run iperf from a VM on the server to a docker container on the server. Or at least more than 1G. So none of the cabling is in play here. Even internally the system is not allowing the 10G speeds while it is acknowledging the 10G is there and it is negotiating that speed with the network. As for the network configuration, I don't use VLANs and its a completely flat network. There are no real settings to actually look at. Everything is allowed to flow. The only thing that COULD be considered different is the DNS server is a VM on the Unraid server. But that would not affect direct IP connections. The only logical issue is going to be the driver I would say. Tracking that change down will be slightly difficult.

I have done this and sadly it did not fix the issue. I read in one post that a driver might have changed that made it happen, but that was from 2019 and this just happened in the last few months.

I am running 6.12.4 and I have a 10G card installed in my server. I also have a 10G card installed on my desktop and was previously able to get 10G to function without issues. Now I cannot seem to make this happen. The server is an HP DL380 G9. It was previously on a Dell R610. Both platforms the 10G worked without issues. My desktop is using a Solarstorm SFN5122F with a 10G capable SFP connected to a fiber run over to the switch which is a PowerConnect 5548P. The Unraid server is using an Intel 10G NIC (forgot model number). The server is connected via a DAC to the switch. I have a speedtest and an iperf3 box setup for testing speeds. Before whatever issue happened I was getting about 6G of the 10G transfer speeds. Enough to know the 10G was working and there were no real issues. I could even get the transfer speeds that showed this to be the case. Without warning and with no changes I can see, suddenly I cant get more than 1G speeds. It is only connected to a 10G NIC, logs show it sees the 10G and connects, Windows VMs see the 10G, the managed switch sees the 10G connected at that speed. So everything should be on the up and up. But no avail. I have tried to use a VM on the machine itself to test the speed to the speedtesting containers and even that doesnt get more than 1G. So even internally without leaving the machine it is only offering 1G speeds. I have set the network bonding to only include the 10G interface I am currently using. All other ports are unplugged so only the 10G should be used in every scenario. I have tested both a docker container and a VM with iperf and the results are the same. Anyone have any ideas that they think might help. A setting, a test, anything. I will share anything needed to help out. This has been driving me mad for about a month. Searches on the forums, reddit and google have not been very helpful. I am hoping a post will get some interest here for someone with new ideas. I have debated on ordering new hardware for the NIC, but I feel this issue is very specific to something not hardware related. It has to be some software trigger. A driver, a setting, an update in the software. Hardware doesn't magically fail from a 10G to a 1G connection perfectly.

I got the container working. However the join command does not seem to work at all in the startup. You have to manually run the join with the networkID to make it work. I checked the connected networks and none were available. But after running the join command it works. So clearly there is something wrong with that part of the container. It works, but certainly not user error when the unraid NETWORKID section is seemingly not even being used.

So there is no plugin designed for just doing KVM. I guess i can use VNC though again not ideal

Just for clarification, i know i can do software kvms but i was looking more toward being a plugin or something within LimeTech's VM host. I dont see it as a very difficult thing to achieve.

So i have been made aware of this software recently and had a question before i sink money into it for the sake of geeking out. I understand this is a VM manager just like any other VM manager (ESXI ...ect) however it is also a bit more configurable (Linux who knew) However, is it possible to use this for two OS on the same box but have a KVM Switch (keyboard, Video, Mouse) where i can use the same mouse and key board on both OS but on like dual monitors..or three. My idea is based on my job. I use Windows but it would be very beneficial to use Linux for some other tasks. As of right now i run a VM on an ESXI host and this isnt ideal due to not having access to it at all times or having to VPN just to reach it and then having to VPN from the linux box to a customer. I have free range over my hardware so this will not be an issue.