https://bz.apache.org/bugzilla/show_bug.cgi?id=65180
--- Comment #6 from Christopher Pereira <krip...@imatronix.com> --- > It means the process is trying to exit. This will happen during graceful > restart, MaxRequestsPerChild/MaxConnectionsPerChild being hit, or due to a > wane in load and MaxSpareThreads being exceeded. > > Looks like your issue is with mod_proxy_http on a long-poll kind of request > as opposed to a websockets tunnel. I think this would rely on mod_proxy > timeouts being able to fire and eventually fail to let the process exit, > unless the backend continually pings data to keep the long-poll connection > alive. Do you know what the data looks like when idle? > > I'd suggest a separate bug if it's not websockets. Yes. CouchDB will be pinging data all the time to keep the long-poll connection alive. So my understanding is that some condition is met and Apache is trying to gracefully finish the thread, but it never finishes it because CouchDB continues pinging data. Is it normal behaviour that long-poll connections are marked as "G" after some time? I will make some tests to confirm that the PouchDB client is still connected and receiving the data, and that the slot is reopened once the client disconnects. In this scenario, I don't see a problem at all. Thanks for clarifying. -- You are receiving this mail because: You are the assignee for the bug. --------------------------------------------------------------------- To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org For additional commands, e-mail: bugs-h...@httpd.apache.org
