CODEWITHSUNDEEP
ruby-on-railsresquemonitrbenv
Fenelon
Fenelon

Reputation: 700

Managing Resque workers with Monit on RBenv setup

I'm trying to set up Monit to manage Resque workers, but it fails to start saying /home/deployer/.rbenv/shims/bundle: line 4: exec: rbenv: not found I've checked that it is running commands as deployer user and if I copy and paste the command directly via SSH everything works fine. Below is my Monit configuration. Thanks!

check process resque_worker_1
  with pidfile CURRENT_PATH/tmp/pids/resque_worker_1.pid
  start program = "/usr/bin/env HOME=/home/deployer RACK_ENV=production PATH=/home/deployer/.rbenv/shims:/usr/local/bin:/usr/local/ruby/bin:/usr/bin:/bin:$PATH /bin/sh -l -c 'cd CURRENT_PATH; bundle exec rake environment resque:work RAILS_ENV=production QUEUE=high,normal,low VERBOSE=1 PIDFILE=CURRENT_PATH/tmp/pids/resque_worker_1.pid >> CURRENT_PATH/log/resque_worker_.log 2>&1'" 
    as uid deployer and gid admin
  stop program = "/bin/sh -c 'cd CURRENT_PATH && kill -9 $(cat tmp/pids/resque_worker_1.pid) && rm -f tmp/pids/resque_worker_1.pid; exit 0;'"
    as uid deployer and gid admin
  if totalmem is greater than 300 MB for 10 cycles then restart  # eating up memory?
  group resque_workers

Upvotes: 0

Views: 2325

Answers (4)

jyothsna gowd
jyothsna gowd

Reputation: 1

Yes, I'm not sure if this helps, but in my monitrc start line, I have to first su to the user I want to run under. I haven't tried to use the uid and gid flags to know if that works well

Upvotes: 0

Oxyless
Oxyless

Reputation: 51

I saw that the topic was created in 2012 but I had a similar problem and this thread is top ranked by google.

The problem is that monit launch commands with a restricted env (env -i PATH=/bin:/usr/bin:/sbin:/usr/sbin /bin/sh to simulate).

To use monit with rbenv you must specify the correct path before your bundle exec command.

PATH=/home/[USER]/.rbenv/bin:/home/[USER]/.rbenv/shims:$PATH bundle exec ...

Example with unicorn:

check process unicorn_dev with pidfile /home/wizville/app/dev.wizville.fr/shared/pids/unicorn.pid
  group dev
  start program =  "/bin/bash -c 'cd /home/wizville/app/dev.wizville.fr/current && PATH=/home/wizville/.rbenv/bin:/home/wizville/.rbenv/shims:$PATH bundle exec unicorn -c config/unicorn.rb -D'" as uid "wizville"
  stop program = "/bin/bash -c 'kill -s QUIT `cat /home/wizville/app/dev.wizville.fr/shared/pids/unicorn.pid`'"
  depends on mysql

Upvotes: 1

hakimkal
hakimkal

Reputation: 90

This worked for me.

check process app_resque_worker with pidfile <%= resque_pid%> start program = "/usr/bin/env HOME=/home/subcaster RACK_ENV=production PATH=/home/subcaster/.rvm/rubies/ruby-2.0.0-p247/bin/ruby:/usr/local/bin:/usr/local/ruby/bin:/usr/bin:/bin:$PATH /bin/sh -l -c \'cd <%= current_path %>; bundle exec rake environment resque:work RAILS_ENV=production BACKGROUND=yes QUEUE=* PIDFILE=<%= resque_pid %>\'"

stop program = "kill -9 cat <%= resque_pid%> && rm -f <%= resque_pid%>"

if totalmem is greater than 2000 MB for 10 cycles then restart

Upvotes: 0

backwardm
backwardm

Reputation: 309

I'm not sure if this helps, but in my monitrc start line, I have to first su to the user I want to run under. I haven't tried to use the uid and gid flags to know if that works well, so this might be a goose-chase of an answer.

I remember having the same issue as you though... everything worked from the command line, but not when monit would do its thing.

For example, in my monitrc, I am monitoring arsendmail with the following:

# arsendmail_rails3
# daemon that watches and sends mail from the rails app
check process ar_sendmail with pidfile /var/www/rak/log/ar_sendmail.pid
start program "/bin/su - mike && /bin/bash -c 'cd /var/www/rak && ar_sendmail_rails3 -b1000 -d -e production'"
stop program "/bin/ps -ef | /bin/grep ar_sendmail_rails3 | /bin/grep -v grep | /usr/bin/awk '{ /usr/bin/print $2}' | /usr/bin/xargs /bin/kill -9"

Upvotes: 1

Related Questions