wes.crockett

Thanks for the reply. That makes sense. I have enabled Host access to custom networks since I don't want to re-port Unraid. That said, I am still getting ERR_CONNECTION_REFUSED when trying to use the proxy server. I added invokeai.domainname.com. When pinging it, i get the IP for the reverse proxy container, but navigating to it doesn't redirect. EDIT: Made some config changes.: Network: Br0 (set to be able to connect to other networks) Set HTTP port to 80 Set HTTPs port to 443 Set domain.com -> DNS server Created A Record for proxy.domain.com -> NPM docker IP Create C-Name for invokeai.domain.com -> proxy.domain.com From the docker container, when I curl invokeai.domain.com I get: curl: (7) Failed to connect to invokeai.domain.com port 80 after 2 ms: Couldn't connect to server Same when I curl proxy.domain.com... It seems like the issue is with how the NPM package is being connected to. EDIT 2: What I noticed after the changes made in the previous edit was that the container was not respecting my port changes. I installed a different copy of NPM from the app store, put in port 80 and 443 and everything worked instantly.

Hey all, I bought a new domain to use for internal-only routing... to prettify urls to internal services. I am using Unifi with a PiHole as DNS server and I set up NGINX Proxy Manager. Publically, I set the A-Record for my domain to be the internal IP of the PiHole server. On PiHole, I created a record for domainname.com -> {Internal IP of NPM}. When I ping my domain internally, I get the NPM servers IP and I can navigate to the NPM server using the domain name. I was able to use NPM to get a cert for *.domainname.com with little issue (a bit of troubleshooting needed.) I set a host entry for watch.domainname.com and set the IP:Port to be for my Jellyfin server. When I navigate to it, I get ERR_CONNECTION_REFUSED though. I'm really at a loss for how to troubleshoot this further. The one thing I am wondering though is: Does NPM and other docker containers need to all be in the same docker network? NPM is in a network where it gets a unique IP on network while all my other containers just use {UnraidIP}:{UniquePort}. Any advice would be greatly appreciated.

To go from 3.0 a0 to 3.0 release, do we need to delete the container, image, and directories?

I'm not sure which branch I'm on. I am using your published container. I ended up having to delete everything except the userfiles directory and then delete the container and readd it.

Getting a new error this week: │ /home/invokeuser/InvokeAI/invokeai/backend/stable_diffusion/diffusers_pipeli │ │ ne.py:39 in <module> │ │ │ │ 36 from invokeai.backend.globals import Globals │ │ 37 │ │ 38 from ..util import CPU_DEVICE, normalize_device │ │ ❱ 39 from .diffusion import ( │ │ 40 │ AttentionMapSaver, │ │ 41 │ InvokeAIDiffuserComponent, │ │ 42 │ PostprocessingSettings, │ │ │ │ /home/invokeuser/InvokeAI/invokeai/backend/stable_diffusion/diffusion/__init │ │ __.py:4 in <module> │ │ │ │ 1 """ │ │ 2 Initialization file for invokeai.models.diffusion │ │ 3 """ │ │ ❱ 4 from .cross_attention_control import InvokeAICrossAttentionMixin │ │ 5 from .cross_attention_map_saving import AttentionMapSaver │ │ 6 from .shared_invokeai_diffusion import InvokeAIDiffuserComponent, Postpr │ │ 7 │ │ │ │ /home/invokeuser/InvokeAI/invokeai/backend/stable_diffusion/diffusion/cross_ │ │ attention_control.py:13 in <module> │ │ │ │ 10 import psutil │ │ 11 import torch │ │ 12 from compel.cross_attention_control import Arguments │ │ ❱ 13 from diffusers.models.attention_processor import AttentionProcessor │ │ 14 from torch import nn │ │ 15 │ │ 16 from ...util import torch_dtype │ ╰──────────────────────────────────────────────────────────────────────────────╯ ModuleNotFoundError: No module named 'diffusers.models.attention_processor' It has a lot before this, but it all seems to be the same. Any idea of a fix for this? Happens almost instantly when starting the container.

Hey all, I have been dinking around a bit and finally had the need to spin up a dedicated VM for a home project. Prior, CA containers have done everything I need. Well, I disabled Docker and VMs to go set my network to enable bridging. Following doing that, though, DNS seemed to cut out. It had been set to automatic. The CA store wouldn't load, for example. From container consoles, and from my VM, I could ping things like google though. Anyways, I set 1.1.1.1 and 1.0.0.1 as my DNS in Unraid and everything began working again. On my router (UDM Pro) I have DNS set to go to my PiHole (not running on the Unraid server). Any easy/general explanation for why Auto DNS wouldn't work anymore? I was planning on going back and setting DNS to my pihole, but I don't want to take everything down again right now... Maybe tonight. Thank you.

That did the trick. Thanks!

Hey! Glad to see this thread is kept up to date. My install broke a month or two ago and I was kind of tired of fighting it... that said, I trashed everything today and started working to get it re-set up. I am encountering the following: ** An error occurred while attempting to initialize the model: "[Errno 2] No such file or directory: '/userfiles/configs/models.yaml'" ** This can be caused by a missing or corrupted models file, and can sometimes be fixed by (re)installing the models. Do you want to run invokeai-configure script to select and/or reinstall models? [y] I saw another post mentioned this and they mentioned creating the folders and restarting the container which did not work. I also validated that my Hugging Face token was pit in. Here are the contents of the ./userfiles directory Any thoughts on approach?

Thanks @JorgeB. I just ran that with success but now I get a NGINX 404 page... Then I realized that I had my HTTPSEverywhere addon enabled! Once I disabled the addon, I was in. I'm wondering why this isn't more predominant as I'm sure a ton of people are going to start running in to this. Thanks again for your help!

Hi all, Last night I went to log in to my Unraid server and it returned a ERR_NAME_NOT_RESOLVED in Chrome. Tried on several machines with the same results. Pinging host name returned valid results to the correct IP. I was able to SSH in as well as access my Docker containers UI pages. Also, file access continues to work on both Windows and from a Linux machine. I checked, and I do see the server listening on :80 and :443 and IP info looks correct. I have attached the diagnostics output to this thread. I am a bit hesitant to reboot as I have seen one thread where a person had similar issues, rebooted, and then arrays failed to start. I do have a Pihole on a separate device (Raspberry Pi). It seems to be functioning fine at this point. Any advice on what to do here? kuiper-diagnostics-20230118-0630.zip

I've been wanting to set this up but am rather new to creating my own docker containers... Any chance for a screen shot or something of how you have this configured?

Thank you for the follow up. My BIOS doesn't have a VT-d setting (just a main 'Intel Virtualization' one.) I'll leave it disabled until a future patch.

yuuuup... makes for a fun Sunday

OK. Cool. Mine did not have VT-d so I will add that to syslinux. Later, I will go back in and enable virtualization in the BIOS. I do plan to do GPU passthrough eventually, so hopefully this does get fixed in a future version.

Hallelujah! Disabling Intel Virtualization did fix the NIC issue and I am able to access my UI via XXXXX.unraid.net. All docker containers up and accessible too (just validating that everything is as it should be.) I have attached my diagnostics (post fix). With Intel Virtualization disabled in the BIOS, I assume I cannot run VMs on this platform. Is that correct? (I'm not currently running any, but just want to make sure I know the limits currently.) Is there a patch planned to fix this issue? kuiper-diagnostics-20220528-1301.zip

Does disabling Intel Virtualization (I can't find anything that is just VT-d) mean I cannot run VM's on this installation? I'm not anyways, just wanting to make sure I know the ramifications.

Working on one of the above attempts now (disabled Intel Virtualization in BIOS). I think that is probably along the lines of the issue. I will upload logs next though for sure. Thanks.

Does this mean i need to take that USB out and plug it in to another system? Or can I copy it to a new USB... i assume i would need to (learn how to) mount the new USB to do this...

Looking in to this now... this looks promising.

The legend Squid has entered the chat It is a Poweredge T420. I ran diagnostics but am not sure how to get them off the box being as how it's CLI only at this point. Simple guide for that somewhere?

Did you ever figure this out? Fighting the same issue now.

Hi all, Just updated to 6.10.2 and, not, my host is unreachable. After stealing a TV from another room, OS seems to boot fine but I cannot get my box to talk to the network. Upon booting, the Unraid login screen (CLI) says IPv4 Not Set and IPv6 Not Set. I checking my network.cfg and it showed DHCP, Bonding, and Bridging as "yes" with no IP, gateway, or mask set. I tried setting all three to no and setting the IP, gateway, and mask manually but was still unable to reach the box. During boot, I did catch the wording: "Device 'eth0' does not exist. Cannot find device 'eth0'" (image attached). Is that because of binding/bridging? I couldn't imagine the process of a reboot borking my NIC... Please let me know what I can try to fix this issue. Thank you, Wes

I've done it with Valheim. Assuming it's the same, for the most part, process to set up, just choose different ports for the second instance (and name it differently.) Set up port forwarding and inform your friends that they may need to include the port number in their connection string.

That's what I thought too. I do see them in my /mnt/user/dockerShares/valheim/Backups folder now, but only starting when the server got rebooted. I may perform a reboot now just to see if they go missing again. Cache option for dockerShares is set to ONLY. That said, I'm wondering if I have a messier issue here as there is a /mnt/cache/dockerShares/valheim/.... folder too. I'm showing my noobness here Thanks for the reply and the thorough run down. I really appreciate it. ---- EDIT ---- As a precautionary method, I set up a Duplicati backup for just the backups folder to copy them out as well. Hopefully that is added insurance.

Just had an issue where I had to reboot my Unraid server. When it came back up, I started up my Valheim container and connected to a completely fresh world... In testing that world, I shut down the container, waited a while and started it back up and it didn't reset. Not sure what to make of that issue. I have backups set to true but I can't find where they are located to restore the old world. Anyone know where they SHOULD be?