2nd relay: docker container, ports, incoming/outgoing

Hi guys,

We have one core and one relay running on one cloud instance each. now we also have a dedicated server where we installed a 2nd relay running as a docker container.

for the cloud instances we followed the coincashew guide. For the 2nd relay we used this guide: https://tdiesler.gitbook.io/cardano/plain-docker/running-the-nodes

Right now our 1st relay runs for 78d and has 30 Incoming 20 Outgoing connections. It is configured just like in the CC guide with port 6000 to the core and 3001 to new-relays domain.

Now on 2nd relay I just confugred the same ports. Toplogoy updater stated “nice you stay with us” but what makes wonder is two things:

A) I only have 2 incoming 9 outgoing connections on 2nd relay
B) When I ‘netstat -n -a | grep 3001’ I am only listening on tcp (+tcp6) while on the 1st cloud relay I have established connections

I configured both CNODE Ports to be 6000. they both are telnetable from outside while 3001 is not. Im really unsure wether everything is working fine for now…


What is your ticker? Is it mainnet or testnet?

I’m not that familiar with their implementation of cardano-node in docker, but I would assume that probably something is not so good with topologyUpdater, maybe you are pushing wrong IP or Port using topology script and that why you are not getting incoming connections? Try to disable it inside docker container and try to run it on host.


Seems a project what build that docker image is archived, so that is a bit alarming:

That seems to imply that no updates will be shipped for that docker image anymore and that means that in 4 weeks after hard fork on mainnet you will need to search for new solution for your new relay. I’m not familiar with that project at all, so maybe I’m wrong and maybe somebody more familiar with that project can comment more on that.

well, seems like I was a little to fast: we gained 10 ins 9 outs meanwhile.

topologyUpdater is integrated and one of the first problems with this was indeed wrong IP. but we managed to fix this issue and now updater says: “glad you’re staying with us”

For the docker image concerns: the last update was 3 months ago. I didnt had to update anything at all when I installed this.

Even if they dont update this anymore I should still be able to upgrade the running cardano node inside docker by myself when hard fork?!

You can try to update cardano-node inside docker, but what about gliveView, cncli and etc. You will need to keep truck of all applications and update them somehow, I would assume that not very easy task… Additionally why to use docker then if everything still need to be done manually?

I personally use official docker image of cardano-node and I think that is a way to go, maybe consider that one.


Looks like that only supports the x86/amd64 architecture? We moved to arm a while back, which nessus-cardano fully supported.

Back to the drawing board, I guess.

Check out Docker Hub. I have a relay and producer running with that image on a k3s cluster.

now that 1.35.3 is released and we have updated two nodes already what is the best practice for me getting another docker instance with 1.35.3 please? the tdiesler image obv is not updated and somehow I struggle to find a good how to for the official docker


If you want to use official IOG docker image, check out my docker-compose file, it should give you some explanation how to run it.

I do plan to create a short manual for running oficial docker containers, but didn’t found time for that.

Let me know if you have any more question, regarding my setup, I would be more then happy to help.