Pulp

When you run pulp_installer 3.10.0 against a very slow machine, such as a Qemu emulated machine with 2 CPU cores (on top of GHA CI, or on my Core i7-8650U system), trying to do any request to the API may result in a "Connection Timed Out" for every single request.

Even the command curl http://127.0.0.1/ would exhibit this.

logs would look like:

Jan 28 20:53:15 pulp3-source-fedora32.lan systemd[1]: Started Pulp WSGI Server.
Jan 28 20:53:19 pulp3-source-fedora32.lan gunicorn[21755]: [2021-01-28 20:53:19 +0000] [21755] [INFO] Starting gunicorn 20.0.4
Jan 28 20:53:19 pulp3-source-fedora32.lan gunicorn[21755]: [2021-01-28 20:53:19 +0000] [21755] [INFO] Listening at: http://127.0.0.1:24817 (21755)
Jan 28 20:53:19 pulp3-source-fedora32.lan gunicorn[21755]: [2021-01-28 20:53:19 +0000] [21755] [INFO] Using worker: sync
Jan 28 20:53:20 pulp3-source-fedora32.lan gunicorn[21774]: [2021-01-28 20:53:20 +0000] [21774] [INFO] Booting worker with pid: 21774
Jan 28 20:53:50 pulp3-source-fedora32.lan gunicorn[21755]: [2021-01-28 20:53:50 +0000] [21755] [CRITICAL] WORKER TIMEOUT (pid:21774)
Jan 28 20:53:50 pulp3-source-fedora32.lan gunicorn[21774]: [2021-01-28 20:53:50 +0000] [21774] [INFO] Worker exiting (pid: 21774)
Jan 28 20:53:51 pulp3-source-fedora32.lan gunicorn[21778]: [2021-01-28 20:53:51 +0000] [21778] [INFO] Booting worker with pid: 21778
Jan 28 20:54:21 pulp3-source-fedora32.lan gunicorn[21755]: [2021-01-28 20:54:21 +0000] [21755] [CRITICAL] WORKER TIMEOUT (pid:21778)
Jan 28 20:54:21 pulp3-source-fedora32.lan gunicorn[21778]: [2021-01-28 20:54:21 +0000] [21778] [INFO] Worker exiting (pid: 21778)

I discovered that by passing to gunicorn (pulpcore-api and pulpcore-content) "--timeout 90" (in the systemd unit files), it fixes this issue. The 1st request may take a while, but later requests would return almost immediately.