sebstrgg

I've had a weird problem for quite a while now. And it's the docker service that kind of "hangs". I only notice in the GUI since the dashboard widget that shows my docker containers doesn't get populated. If I then try to go to the docker page, it just loads forever. The good, but still annoying thing with this is that all dockers continues to operate as normal. BUT, I can't do anything with the docker service. Any CLI command just hangs if I try to manually control anything docker related. Also I'm totally unable to do a graceful shutdown or reboot of the machine since unraid also can't shutdown the docker service. I think I've tried pretty much everything that comes to mind now which includes turning on/off container js option, recreating docker image, rebooting (god knows how many times) and upgrading. After I've done a forced reboot and get back into the Unraid GUI everything works as expected for a few hours, somtimes a few days - but it always ends up back in the same situation as mentioned above. The problem probably started in a late 6.X stable version and I'm currently on 7.0-beta2. Really nothing helps.

To be able to use a container as the network for other containers, you need to use the 'docker network create container:<container_name>' command as per other guides. No changes there. The problem is that when the VPN container is restarted in some way (connection restarted or container) the containers connected to that network most often drops their connection overall. If the container is restarted it's due to the VPN container (master) changed container ID, leaving the connected containers hanging in wait of a restart. Here you can use the Rebuild-DNDC container to get help to restart the connected containers. The problem I've experienced is that the Rebuild-DNDC not always manages to properly restart the containers, or that when it's been a connection reset of any kind (leaving the master container intact with the same container ID), it fails to recognize this and the connected containers also drop their outbound connection (to the internet). I'm sure Rebuild-DNDC works properly if used with gluetun as it's built for that purpose - but since I don't, this is an annoying problem leaving me to manually restarting the containers.

Hi all, I'm not sure this is the right part of the forum to post, so mod, please move in case you find a more suitable place. I just wanted to share my solution on the topic which I hope can help someone else! My problem: I want to utilize a container with a VPN client as a docker network, to enable chosen containers to always be connected through the VPN connection. I tried all possible solutions found here and other places to make this work over time, but I always ended up having problems with having to manually restart these containers. I've tried the Rebuild-DNDC container on multiple occassions, but as it's primarily made to be use in combination with gluetun, and I unfortunately didn't get it to catch soft failures of containers properly. Solution: I put my (very) limited code production skills to use with the help of AI and created a bash script to be used with a cronjob with the User Scripts plugin. I wanted to share my now (for me) confirmed working solution as it's been running for two weeks straight without any hiccups. You can find the script here: https://github.com/sebstrgg/unraid-restart-vpn-containers I'm open to any suggestions, feedbacks, pull requests etc. to make the script more user friendly and resilient. There might be obvious sub-optimizations as I don't have that much skill in coding overall and AI made a majority of the code with a lot of revisions through prompts from my part.

I actually discovered this late last night going through the iptables of the different dockers. Great work anyhow, your dockers are actually really impressive! Since you're already gooing around the iptables ruleset, is there any chance to add the ability to add custom iptables rows? The reason I see a need for this is due to the fact that running plex behind a vpn is troublesome unless you're allowed to setup NAT to and from a specific port (a open port in your vpn -> plex port). This works perfectly when entering the line manually and the problem is that I don't see a way to apply it automatically upon start of container/connection to vpn since your startup-script clears the .ovpn config from any potentially conflicing lines. The iptables rule I'm using to make this work is: iptables -t nat -I PREROUTING -p tcp --dport [forwarded vpn port] -j REDIRECT --to-ports [plex port, i.e. 32400]

Hey, Since I'm routing other dockers network via this docker, I'd like to do an addition to the iptables rule list (nat) pointing from an open external port via my vpn provider to a port attached to the docker. However, I'm unable to do so as the docker automatically clears any changes done to for example the .ovpn file (tried adding "up" command). Is there any way to add an iptables row to be automatically applied on startup/connection to vpn?

I get the exact some problem. I've been trying to route multiple docker containers (firefox, xteve, hydra2) via binhex-delugevpn and binhex-sabnzbdvpn. Although I've done everything exactly as specified, I still end up with the dockers that are being routed not being accessible over my local network, but I can verify that the dockers are live and functioning towards the internet via the vpn connection. I did notice a strange behaviour when checking open ports on the vpn-container (netstat -plnt) to verify that the ports were open, and sometimes they are, sometimes they aren't and sometimes they only seem to map up over ipv6. I really don't know enough about docker- and unraid-specific docker networking to troubleshoot it any further so any help would be greatly appreciated. EDIT: I was running Unraid 6.8.3-stable with this behaviour and updated to 6.9.0-beta1 to verify that the problem is still occuring, which is is.

Yes you're right, but "move" is not the problem, I need to "copy" completed files.

I'm essentially only trying to achieve what the plugin does - copy a completed file to another folder. Nothing special tbh

Hey, I'm trying to get the Copy Completed plugin working, but as far as I've gathered right now, I understand that there's an issue with the Python version and the plugin mainly being unmaintained. (The plugin only exists for python2.6, python2.7 while the container uses python 3.7) Is there anyone that have this plugin working? If so, how did you get it working? If not, do you use any other plugin to achieve the same end goal? Thanks /S

Not a problem at all! I'm happy to help You're perfectly right (or rather the other user) and the inital problems with the containers is officially fixed! I got it working right away! Thank you!

Ok, so I just re-did the procedure and ended up with the same error as before, but now with an additional error. Full Run log: https://wl7r.me/KIBh8R (This was tested just now, after your second update if I understood the previous posts correctly)

Hi again, My setup is identical to yours, except for the IS_CREATION_ENABLED and CONNECT_WITH_FRANZ variables set to true. The reason I did try to add PUID/PGID was only due to the error message I ended up with and I've already fully removed the container, appdata-folder and re-added the container. I could try it again if you believe it would make any difference. Run log (with modified PUID/PGID): https://wl7r.me/TIMnQA Run log (as per standard container setup): https://wl7r.me/iRCmDH Screenshot of container settings: https://wl7r.me/uLT2ou EDIT: So I did it again, I removed the container + image and deleted the appdata-folder agian. I just changed the DB string to correspond with my DB-server. New run log: https://wl7r.me/nkFjoD I don't know if its relevant, but I do run Unraid v6.8.0 rc1, which have a newer docker version. EDIT2: I retried the recreation of the container as I noticed the search path for /config was "/mnt/user/appdata" instead of "/mnt/cache/appdata" like on the the /recipes and /database. No change in the error type though. EDIT3: Same error when changing search path for /config, /recipes, /database to "/mnt/user/appdata" in the beginning of the path. EDIT4: I use your container that is provided in the Unraid CA repo (https://wl7r.me/EJaMD4)

Hi, I've just tried to setup this docker container, which seems to fail based on the user permissions for NPM. npm ERR! code EUIDLOOKUP npm ERR! lifecycle could not get uid/gid npm ERR! lifecycle [ 'nobody', 0 ] npm ERR! lifecycle I've tried to set the container variables for PUID/PGID to 99/100 without any luck. Any suggestions on how to proceed? Thanks, Sebastian

Hi, Thanks for your reply! 1) Ok, so what you're saying is that my letsencrypt docker which have the following settings: custom network: proxynet (172.18.0.0/16) -> which automatically gets bridged to host_ip:180/1443, won't be able to communicate with anything on my lan via the host ip, but only other dockers connected to the same internal network of the docker host? Is there any way I can make it possible to use my letsencrypter-docker for ALL things I'd like to proxy then? I'm not 100% sure of what you mean by the macvlan, but I suppose you're referring to the internal lan inside the docker host?

Hi, I've just started using Docker with my transition to Unraid as I've up until now have been running the VMware suite, "full-VM setup" and a self-configured nginx reverse proxy setup for my public services. As I haven't used Docker before, there's probably something obvious I'm missing, but here goes; Two main problems/questions: 1. In the case where I'd like to proxy a docker that's NOT on the 'proxynet' but instead connected to one of my actual LAN-networks (i.e. VLAN50/subnet: 10.10.50.0/24), I'm not able to do this successfully. I've made sure that I've made the corresponding changes to the nginx proxy conf (i.e. home-assistant docker) and its 'upstream_address' (<docker-name> -> <local IP>). I'm able to access it via the <host_IP>:<docker_port>, and even though this is a match in the config-file, It still doesn't pass through the traffic properly (bad gateway). Any ideas of what I might be missing out on? (I've tried this with several docker-containers unsuccessfully). This is also highly relevant as I want to be able to proxy 'anything' with the letsencrypt-docker 2. How do you (others) have pfsense setup to port forward specific traffic, via a specific gateway to a docker-container with the 'proxynet' setup? (see my last bullet point below) Some background info: - letsencrypt-docker is confirmed working and is working very well - I configured it with the help of SpaceinvaderOne's Youtube-guide and the 'proxynet' setup - I use sub-domains and not sub-folders - All Dockers that are connected to my 'proxynet' are successfully being proxied - I have a fairly advanced local network which is segmented into different vlans - I run a pfsense instance with an active OpenVPN-client connected to a VPN-service, where I've excluded specific VM-traffic via changing the gateway specific traffic is running over, when the need have arised to use my public IP. I might've missed some crucial info that would help you to help me, but just let me know what you need if that's the case. There's a lot of new stuff I'm poking around with (unraid and docker) and although there's been some hiccups such as this, it's still a lovely experience. Great job with the letsencrypt-container, it's truly awesome!