Rootless podman 'Error: pasta failed with exit code 1: Failed to bind port 8765 (Address already in use) for option -t 8765-8765:3000-3000' (exit 126) when re-running a dashboard container
Symptom: re-running a Next.js dashboard container for a Nessie project with podman run --rm -d --name nessie-dashboard -p 8765:3000 ... failed immediately with 'Error: pasta failed with exit code 1: Failed to bind port 8765 (Address already in use) for option '-t 8765-8765:3000-3000'' and exit code 126. No application logs were produced.
Cause: rootless podman (4.4+, the default on recent Debian/Fedora) forwards published ports with pasta, which binds the host port itself. If anything already holds that port, the container never starts. Usually that's a previous detached container from an earlier run, a next dev server, or another project's service. The error comes from the network helper, so it doesn't look like the usual 'address already in use' from your app.
Fix: find and free the port, or pick another one:
podman ps --format '{{.Names}} {{.Ports}}' | grep 8765
ss -ltnp | grep ':8765'
podman rm -f nessie-dashboard # if it is your old containerTo make re-runs idempotent, add --replace together with --name (podman replaces an existing container of the same name). Choosing a different host port also works (we moved to -p 9876:3000), but update anything that hardcodes the URL, such as Playwright's baseURL or a webhook/tunnel target. With several containers in one hackathon project, write the ports down in the README to avoid collisions.
Verify: curl -s -o /dev/null -w '%{http_code}' http://127.0.0.1:<port>/ returns 200 after the container starts.