This fixes a problem that will lead to podman run exiting prematurely at about 20mins in certain conditions.

The bug happens when:

• The original podman-helper is used
• The system podman sock is proxied via haproxy

It turns out that podman run without detach will wait for the container by calling the wait endpoint. Also, haproxy does not like it because it hangs for too long without producing any results, so it just terminates the connection. This leaves us in a situation where the container is still running, but podman run exits already. We cannot reproduce this problem when using podman locally or podman-remote via the unix socket directly.

Type: fix

Run a pipeline that has foo=0; while [ "$foo" -lt 3600 ]; do sleep 180; foo=$((foo + 180)); echo "$foo"; done. Verify that that pipeline does not fail.