On Mon, Feb 28, 2011 at 10:04 PM, Renato Isidio <renato...@gmail.com> wrote:
> Hi Marius, > > today we saw that our git-daemon was down. The message follows: > > /opt/gitorious/script/git-daemon:76:in `peeraddr': Transport endpoint > is not connected - getpeername(2) (Errno::ENOTCONN) > from /opt/gitorious/script/git-daemon:76:in `run_service' > from /opt/gitorious/script/git-daemon:70:in `run' > from /opt/gitorious/script/git-daemon:52:in `start' > from /opt/gitorious/script/git-daemon:321 > So it seems the connection is lost, and the accept_socket isn't working. Could it be the server is on a not-so-reliable network connection? These things can happen, I suppose. One thing you could do is to set up some kind of process monitoring on your server. We use Monit ( http://mmonit.com/monit/) on gitorious.org, and forcefully close services that aren't performing as suspected. Most distros will have a Monit package, and the defaults will usually be just fine. What you could do is to have Monit monitor your git daemon processes, it will do this based on a pid file. In the Monit "recipe" you could add a check that verifies that your git-daemon is available on the specified port (9418); if not restart the process. It would look something like this: check git-daemon with pidfile /var/www/gitorious/tmp/pids/git-daemon.pid start program = "/etc/init.d/git-daemon start" stop program = "/etc/init.d/git-daemon stop" if failed port 9418 then restart Cheers, - Marius -- To post to this group, send email to gitorious@googlegroups.com To unsubscribe from this group, send email to gitorious+unsubscr...@googlegroups.com