Mongrel PID files and reboot

#1

I have followed the instructions on the mongrel website, and now have
mongrel_cluster running my apps at boot time, works great.

Now, if the server goes down ungracefully, the mongrel processes do
not start up on reboot because the mongrel PID files are still hanging
around.

I have boot job that cleans those up now, but is there a “GOOD” way to
handle this ???

– Thanks

#2

Well, that’s how I’ve handled it… I’d also like to know if anyone has
any
better methods.

#3

Actually, I just found this, its looks cleaner than what I have…

http://textsnippets.com/posts/show/931

Basically just a change to the mongrel_rails script to delete PID
files that reference a process that is not actually running…sounds
like a good thing to be included to me.

#4

this is what I am using now (changed in mongrel-1.0.1/bin/
mongrel_rails
if File.exist? defaults[:pid_file]
#===== new stuff
pid = File.new(defaults[:pid_file]).readline
if ps --pid #{pid} --no-headers.length > 0
log “!!! PID file #{defaults[:pid_file]} exists, but
is stale, and will be deleted so that this mongrel can run.”
File.delete(defaults[:pid_file])
else
log “!!! PID file #{defaults[:pid_file]} already
exists and the process id referred to in it is running. This mongrel
is probably already running. #{defaults[:log_file]} for errors.
EXITING.”
exit 1
end
#===== old stuff
#log “!!! PID file #{defaults[:pid_file]} already exists.
Mongrel could be running already. Check your #{defaults[:log_file]}
for errors.”
#log “!!! Exiting with error. You must stop mongrel and
clear the .pid before I’ll attempt a start.”
exit 1
end

#5

oops, this is actually what I am running now (again patched into /usr/
lib/ruby/gems/1.8/gems/mongrel-1.0.1/bin/mongrel_rails)

      if File.exist? defaults[:pid_file]
        #===== new stuff
        pid = File.new(defaults[:pid_file]).readline
        if `ps --pid #{pid} --no-headers`.length > 0
            log "!!! PID file #{defaults[:pid_file]} already

exists and the process id referenced is running. This mongrel is
probably already running. #{defaults[:log_file]} for errors.
EXITING."
exit 1
else
log “!!! PID file #{defaults[:pid_file]} exists, but
is stale, and will be deleted so that this mongrel can run.”
File.delete(defaults[:pid_file])
end
#===== old stuff
#log “!!! PID file #{defaults[:pid_file]} already exists.
Mongrel could be running already. Check your #{defaults[:log_file]}
for errors.”
#log “!!! Exiting with error. You must stop mongrel and
clear the .pid before I’ll attempt a start.”
end

#6

Found this thread searching the archives, after running into this same
problem. Anyone know why this hasn’t been added to mongrel_cluster trunk
yet? Very annoying. Or maybe it has been, but just wasn’t in the distro
I got?

Problem is if your server/mongrels dies ungracefully, then
/etc/init.d/mongrel_cluster won’t succesfully start up your mongrels
again, since there are stale pids.

removed_email_address@domain.invalid wrote:

oops, this is actually what I am running now (again patched into /usr/
lib/ruby/gems/1.8/gems/mongrel-1.0.1/bin/mongrel_rails)

      if File.exist? defaults[:pid_file]
        #===== new stuff
        pid = File.new(defaults[:pid_file]).readline
        if `ps --pid #{pid} --no-headers`.length > 0
            log "!!! PID file #{defaults[:pid_file]} already

exists and the process id referenced is running. This mongrel is
probably already running. #{defaults[:log_file]} for errors.
EXITING."
exit 1
else
log “!!! PID file #{defaults[:pid_file]} exists, but
is stale, and will be deleted so that this mongrel can run.”
File.delete(defaults[:pid_file])
end
#===== old stuff
#log “!!! PID file #{defaults[:pid_file]} already exists.
Mongrel could be running already. Check your #{defaults[:log_file]}
for errors.”
#log “!!! Exiting with error. You must stop mongrel and
clear the .pid before I’ll attempt a start.”
end

#7

Okay, I found that, at least in the newest versions of mongrel_cluster,
you can supply a --clean argument to have mongrel_cluster delete any
leftover pids.

Anyone know how I get this to happen on boot? Right now I’m just linking
/etc/init.d/mongrel_cluster to
mongrel_cluster_gem/resources/mongrel_cluster . Any way short of hacking
the source (meaning it will break next time I update) to get --clean to
be the default for mongrel_cluster start?

Jonathan

Jonathan R. wrote:

Found this thread searching the archives, after running into this same
problem. Anyone know why this hasn’t been added to mongrel_cluster trunk
yet? Very annoying. Or maybe it has been, but just wasn’t in the distro
I got?

Problem is if your server/mongrels dies ungracefully, then
/etc/init.d/mongrel_cluster won’t succesfully start up your mongrels
again, since there are stale pids.

removed_email_address@domain.invalid wrote:

oops, this is actually what I am running now (again patched into /usr/
lib/ruby/gems/1.8/gems/mongrel-1.0.1/bin/mongrel_rails)

      if File.exist? defaults[:pid_file]
        #===== new stuff
        pid = File.new(defaults[:pid_file]).readline
        if `ps --pid #{pid} --no-headers`.length > 0
            log "!!! PID file #{defaults[:pid_file]} already

exists and the process id referenced is running. This mongrel is
probably already running. #{defaults[:log_file]} for errors.
EXITING."
exit 1
else
log “!!! PID file #{defaults[:pid_file]} exists, but
is stale, and will be deleted so that this mongrel can run.”
File.delete(defaults[:pid_file])
end
#===== old stuff
#log “!!! PID file #{defaults[:pid_file]} already exists.
Mongrel could be running already. Check your #{defaults[:log_file]}
for errors.”
#log “!!! Exiting with error. You must stop mongrel and
clear the .pid before I’ll attempt a start.”
end