listen tcp :5000: bind: address already in use\n

[dnielsen]

I ran docker compose up -d, which seemed to download and build everything correctly. But then I get the following error:

Error response from daemon: something went wrong with the request: "listen tcp :5000: bind: address already in use\n"
Error: executing /usr/local/bin/docker-compose up -d: exit status 1

When I run lsof I don't see any pid at 5000.

Google     4891 dnielsen  556      PIPE 0xdb68501d4e3c9b7c        16384                     
Google     5005 dnielsen  cwd       DIR               1,17          704                   2 /

I did get the following warnings:
WARN[1322] Docker Compose is configured to build using Bake, but buildkit isn't enabled

Setting up fakeroot (1.31-1.2) ...
update-alternatives: using /usr/bin/fakeroot-sysv to provide /usr/bin/fakeroot (fakeroot) in auto mode
update-alternatives: warning: skip creation of /usr/share/man/man1/fakeroot.1.gz because associated file /usr/share/man/man1/fakeroot-sysv.1.gz (of link group fakeroot) doesn't exist

update-alternatives: using /usr/bin/xz to provide /usr/bin/lzma (lzma) in auto mode
update-alternatives: warning: skip creation of /usr/share/man/man1/lzma.1.gz because associated file /usr/share/man/man1/xz.1.gz (of link group lzma) doesn't exist

Note: I am using docker via podman-docker.

[deanwampler]

I saw this after stopping and starting podman machine several times yesterday. I never figured out the cause, but my theory is that when I ran podman machine start at some point, it restarted the llama-stack container it ran before. Not sure. It went away when I rebooted my machine :)

[nikhilmaddirala]

Thanks @dnielsen - I suspect there might be two separate issues here:

1. Incorrect port mapping set up

2. Incompatability with podman

Regarding (1) - can you try updating your .env file to use a different port and try it with docker instead of podman?

Regarding (2) - I'm not very familiar with podman so this needs some investigation

[deanwampler]

I suspect podman, too. FYI, based on your suggestion yesterday, I removed the health checks to see if they were causing the crash loops, but they still happened. 🤷

[deanwampler]

Some further observations.

1. Install Podman from the Podman.io site. It has newer features than the brew version
2. I am currently observing OOM errors in the llama-stack container. Since Podman has you run a VM, upon which everything is executed (at least on a Mac...), I need to investigate how to give it more memory.
3. I have seen the "ping port 5000" health check fail, so there may be network issues. I'll investigate once I solve the memory problem.

I'll also try using docker on my personal Mac to confirm it works okay. I'm also going to add a ticket for creating an alternative deployment that just runs the processes individually (though less conveniently...).

[deanwampler]

On the original issue topic, I think it probably also helps to be sure to use docker compose down when you want to restart, rather than "nuke" the VM 😉

[deanwampler]

@dnielsen have you been able to run the app? Note that I just merged an updated README.md with instructions for running the services individually, for situations where docker isn't working (e.g., trying to use podman compose). Let us know if you want to keep this ticket open.