Install Cloudron 7.0.3 on Ubuntu 20.04.3 ESXi

[girish]

@mannyc1701 Can you check if docker is running ? systemctl status docker and then docker network inspect cloudron should print a bunch of 172.18.x.x addresses.

If that is working:

• lsof -i :53 - this will print who is listening on port 53. Sometimes, systemd-resolved or bind9 (named) is enabled by default in some images preventing unbound from running. We actually disable this in our setup scripts, but helps to double check.

[mannyc1701]

@girish Docker Status:

Docker Network & port 53:

Thank you!

[girish]

@mannyc1701 docker seems OK. What is the lsof -i :53 output?

[mannyc1701]

@girish Its blank, no output.

[girish]

@mannyc1701 Does nc -lu 172.18.0.1 53 work ? This tries to listen on UDP port 53. If it works, "it will just hang". Otherwise, it will print nc: Address already in use or something.

And also nc -l 172.18.0.1 3003 ? This tries to listen on TCP port 3003.

[girish]

@mannyc1701 also, have you tried rebooting the VM just in case and check if those services start up afterwards ? thanks!

[mannyc1701]

@girish both show the same error, cannot assign requested address.

The box screenshot mentions this error too, address 172.18.0.1 as not available.

The local server network is 172.17.1.0/24.

[girish]

@mannyc1701 I can't see anything obvious. 172.18.0.0 is the internal docker network (internal to the VM). For some reason, services are unable to listen on that. Are other docker things running in docker ps ? Maybe a VM reboot helps, not sure.

Out of curiosity, you can also do:

docker run -ti cloudron/base:3.0.0 /bin/bash and then in the shell there, nc -l 0.0.0.0 3000 does that work?

[mannyc1701]

@girish here are the outputs:

Prior to reboot after installation I ran ip addr output and could ping 172.18.0.1 & 172.18.0.2.

this is the output after restart.

docker ip changed to 172.19.0.1, I don't if that helps, it seems ip 172.18.0.1 is in use or configured and cannot be used but I can reach 172.18.0.2 without issue.

[girish]

@mannyc1701 said in Install Cloudron 7.0.3 on Ubuntu 20.04.3 ESXi:

cannot be used but I can reach 172.18.0.2 without issue

Thanks for hanging in there. The above is the core issue. Our code assumes the bridge gets the IP 172.18.0.1 . I am not sure why it gets it as 172.18.0.2 in ESXi . Maybe it was just a happy accident, but we had this assumption since forever.

Can you try this then:

• First, remove all existing containers: docker ps -aq | xargs -r docker rm -f
• Delete existing cloudron network: docker network rm cloudron
• Recreate cloudron network for the IP: docker network create --subnet=172.18.0.0/16 --ip-range=172.18.0.0/20 --gateway 172.18.0.1 cloudron
• Make cloudron code recreate containers: sed -e 's/48.20.0/48.19.0/' -i /home/yellowtent/platformdata/INFRA_VERSION
• Does systemctl restart unbound work after above? If it does, you can also do systemctl restart box which should also hopefully work.

[mannyc1701]

@girish That did it, Thank you so much for helping me get this going!

[girish]

@mannyc1701 thanks, I pushed a fix as well for next release - https://git.cloudron.io/cloudron/box/-/commit/fdefc780b4f687ec5ff508b1a37270492612f2a3