we are running a ubuntu machine as our demo server for serving Ruby on
rails apps
we are using mysql,nginx,mongrel for serving the sites.
the problem is if for any reason the power fails out and comes back
mysql and nginx get started fine but mongrel does not work due to
some pid(mongrel.8000.pid mongrel.8001.pid mongrel.8002.pid) files
existing in log folder.
we then manually delete the files and start the cluster.
is there anyway we can check during reboot of ubuntu server for any
existing pid
files in log folder and delete them and then start the mongrel cluster.
the problem is if for any reason the power fails out and comes back
sounds like the first thing to do is get a UPS
mysql and nginx get started fine but mongrel does not work due to
some pid(mongrel.8000.pid mongrel.8001.pid mongrel.8002.pid) files
is there anyway we can check during reboot of ubuntu server for any
existing pid
files in log folder and delete them and then start the mongrel cluster.
Of course – look in /etc/init.d for examples of scripts that are run at
boot time (and/or shutdown).
Write your own and link it to the appropriate run-level (rc*.d)
directory.
Ubuntu and debian mount /var/run/ directories using tmpfs, which means
the
/var/run directories are cleared out at boot time. That’s why debian
packages write out pid files into subdirectories of /var/run. You might
be
able to configure your mongrel so that the pid files are written under
/var/run
