OpenCore local login page not reachable after fresh install using docker compose files

OpenCore
1

Hi,

I’m trying to install OpenCore locally using the docker compose files from git. After an succesfull installation, everything seems to be up an running in Docker. However, when opening the login page using http://localhost.openiap.io, I get an connection refused error in Chrome. Pinging the adress, I got an answer for 127.0.0.1 as it supposed to be.

In Docker, I get lots of those error messages for traefik: “level=error msg="Failed to retrieve information of the docker client and server host: Error response from daemon: " providerName=docker”

Does it have something to do with my docker setup? I tried on two machines with freshly installed Docker, both show the same results.

Kindest Regards

Patrick

2

This looks like traefik cannot reach the docker daemon. Did you edit the docker-compose file? You need to preserve this line in the traefik config:

    volumes:
      - "//var/run/docker.sock:/var/run/docker.sock:ro"

If you are on linux, also make sure you have permission to the docker daemon by running:

ls -la /var/run/docker.sock
3

Thansk for quick answer. I’m running Docker on Windows (it is using WSL) and did not modify the docker-compose file. The permissions are for root only. At least that’s what ls -la /var/run/docker.sock told me. Perhaps, the issue comes from WSL. I’m new to WSL and don’t know much about. But I’m quite familiar to Linux. I may give Docker a try there.

4

I know many people are running this on Docker Desktop on Windows, using WSL, so that should definitely be possible.

I had a different issue a few days ago, where someone was using a Docker-compliant alternative, and it had issues with the old version of Traefik the Dockerfile uses. So you could also try removing the version tag from image: traefik:v2.10.6 to image: traefik:latest and then make sure you uncomment the lines to keep gRPC working without constant disconnects:

      # - "--entrypoints.web.transport.respondingTimeouts.writeTimeout=90000s"
      # - "--entrypoints.web.transport.lifeCycle.requestAcceptGraceTimeout=90000s"

to

      - "--entrypoints.web.transport.respondingTimeouts.writeTimeout=90000s"
      - "--entrypoints.web.transport.lifeCycle.requestAcceptGraceTimeout=90000s"
5

I applied the changes you’ve proposed and at least I don’t see any error messages for traefik in Docker terminal anymore. For some reason, after stopping and restarting the api in Docker, I was able to reach the login and could do a login and the admin user was created. However, soon after login it did not work anymore. Another try to login and it again refuses the connection. So, after some brainless restarting of the components and testing subsequently I could get in again, open all the pages from the left side, but I couldn’t create anything.

6

So, now I seem to have figured out, that a restart of rabbitmq seem to help. No I’m able to create agents and so on. I’ve no clue, what’s actually going on and why the restart helps. Let’s see how long it runs

7

i would try going back to using image: traefik:v2.10.6 if you can make that work with a few restarts, that is what the system was tested on.

In that setup, the only known issue, is mongodb needs a little time to create the replicate set, that is why you get a message saying you need to run docker-compose up two times, the first time.

8

I have gone back to v2.10.6. The error on traefik was then back again. So, going to latest seem at least to solve that issue. I found out it is better to wait for a few minutes before first connection. In the meantime, I was able to set up a nodered agent in Openflow. The strange thing now is, that nodered looses connection every 1:30 to 2 min with messages in api.

Example:

[server][INF] close rnl0mlkft grpc 172.20.0.2 undefined

[server][INF] Disconnected client, client count 3

[server][INF] Client connected, client count 4

[server][INF] Client connected, client count 5

[server][INF] Notify patrick of replace settings for empty-hill-2bde

[server][INF] close hoy4yzuec grpc 172.20.0.2 undefined

[server][INF] close hoy4yzuec grpc 172.20.0.2 undefined

[server][INF] close r0989jjtu grpc 172.20.0.2 undefined

end

9

Yeah, that is the problem traefik introduced around 2.7 and we want to stay at 2.6 until they fix it or find a work around. Funny thing is, there is no problems when using kubernetes or openshift

10

Okay. Thats good to know. I will try different versions of traefik to find if I can solve the problem that way. In the meantime I will dive into OpenRPA standalone.

Thanks a lot for your help, Allan, and have a nice weekend.

Cheers

Patrick