[Dancer-users] many dancer apps with Starman (part 3)

David Precious davidp at preshweb.co.uk
Mon Jul 18 23:56:13 CEST 2011


On Monday 18 July 2011 22:43:25 Mr. Puneet Kishor wrote:
> The above is great help. However, I am trying to use `start_server` with
> `plackup` options and getting the following error
> 
> $start_server --port=5000 --pid-file=pids/app.pid
> --status-file=status/app.status -- plackup -s Starman -w 10 -E production
> -D -a ~/Sites/blog/bin/app.pl start_server (pid:4453) starting now...
> starting new worker 4454
> new worker 4454 seems to have failed to start, exit status:0
> starting new worker 4461
> new worker 4461 seems to have failed to start, exit status:0
> starting new worker 4468
> new worker 4468 seems to have failed to start, exit status:0
> starting new worker 4475
> ^Cworker 4475 died unexpectedly with status:0, restarting
> starting new worker 4482
> received INT, sending TERM to all workers:4482
> worker 4482 died, status:0
> exitting
> 
> and, in spite of the above, it seemed to have started several starman
> master processes. That is not what I wanted.
> 
> $ps -ax | grep 'starman master'
>  4455 ??         0:00.01 starman master           
>  4462 ??         0:00.01 starman master           
>  4469 ??         0:00.01 starman master           
>  4476 ??         0:00.01 starman master           
>  4483 ??         0:00.01 starman master           
>  4496 ttys001    0:00.00 grep starman master
> 
> of course, the bare plackup command works fine
> 
> $plackup -s Starman -w 10 -E production -D -a ~/Sites/blog/bin/app.pl
> 
> The above starts only one master and 10 workers.
> 
> Suggestions?

I suspect that start_server is expecting what it runs to stay around, whereas 
plackup forks a new process then exits - and every time it does, start_server 
tries to start another.


-- 
David Precious  ("bigpresh")
http://www.preshweb.co.uk/

   "Programming is like sex. One mistake and you have to support
   it for the rest of your life". (Michael Sinz)


More information about the Dancer-users mailing list