)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":38496,"name":"Andressa Cabistani","display_name":"Andressa","email":"acabistani@gmail.com","username":"andressadotpy","status":"I\u0027m a Software Engineer at Red Hat and I love Open Source and connect with people! Feel free to DM through IRC, I\u0027ll be delighted to chat"},"change_message_id":"557350d0b1f9447f43163c98f2bd8e272d5a84a3","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":5,"id":"edaa859c_0f872f69","updated":"2026-07-15 11:58:58.000000000","message":"This patch is really good and well-designed.\n\n- The signal mapping in `signal_for()` is correct for both modes, I verified against gunicorn\u0027s signal docs.\n- The warning messages are excellent, really helpful for the operator\n\nJust a small suggestion from one thing that I noticed from my findings: when all target pids are dead (stale pid files from an out-of-band kill) `_target_uses_eventlet()` returns `None` because `/proc/\u003cpid\u003e/environ` is gone, and `_action_signal()` prints a warning like:\n\n```\nWarning: could not determine whether proxy-server runs eventlet or gunicorn; sending SIGTERM, which stops either but skips eventlet\u0027s graceful drain. Pin USE_EVENTLET in the server environment to avoid this.\nSignal proxy-server  pid: 1780  signal: 15\nSignal proxy-server  pid: 1781  signal: 15\nNo proxy-server running\nStarting proxy-server...(/etc/swift/proxy-server/proxy-noauth.conf.d)\nStarting proxy-server...(/etc/swift/proxy-server/proxy-server.conf.d)\n```\n  \nReproduced with `USE_EVENTLET\u003dfalse swift-init proxy-server stop --graceful` after killing the processes out-of-band.\n\nThe behaviour is correct (SIGTERM is safe) but the message is confusing because it implies ambiguity about a running server\u0027s mode when the processes are already dead. The suggestion is to suppress the warning when no target processes are actually reachable, since there\u0027s nothing to missignal.","commit_id":"4e29de05e0e0e1f7be2b642df5390d571dda34be"}]}