TL;DR

fixes supabase start and supabase stop failing every few runs on podman machine

whats hurting the user?

on macOS with podman machine and no docker on PATH, supabase start fails every few runs with failed to inspect docker image and ssh: handshake failed: EOF

supabase stop fails the same way with failed to stop container: exit 125 and leaves the stack behind

both fire one podman call per image or container all at once, and each call opens its own ssh session into the VM, whose sshd drops new ones once ten are still in the handshake

now fixed by?

capping those calls at 8 in flight so every handshake gets accepted, rather than retrying, which would only hide the drops and add delay

flowchart LR
  subgraph before["before"]
    A["start / stop"] --> B["one podman call per image or container, all at once"]
    B --> C["sshd drops past ten handshakes"]
    C --> D["start / stop fail"]
  end
  subgraph now["now"]
    E["start / stop"] --> F["at most 8 podman calls in flight"]
    F --> G["all handshakes accepted"]
  end

ref:

• broken/missed in: feat(cli): port supabase start command to native TypeScript #5847
• spotted while working on: fix(cli): stop start failing on podman without snippets (CLI-2690) #7089