No PID file error deploying with sunspot with Capistrano

Question

I am trying to deploy Sunspot Solr with Capistrano. I have been setting this up based on this gist: https://gist.github.com/doitian/1795439.

deploy.rb

namespace :deploy do
  task :start do ; end
  task :stop do ; end
  task :restart, :roles => :app, :except => { :no_release => true } do
    run "#{try_sudo} touch #{File.join(current_path,'tmp','restart.txt')}"
  end

  desc "Migrate Database"
  task :migrate_db do
    run "cd #{current_path} && rake db:migrate RAILS_ENV=production"
    run "touch #{current_path}/tmp/restart.txt"
  end

  desc "Create Solr Directory"
  task :setup_solr_data_dir do
    run "mkdir -p #{shared_path}/solr/data"
  end
end

namespace :solr do
  desc "start solr"
  task :start, :roles => :app, :except => { :no_release => true } do
    run "cd #{current_path} && RAILS_ENV=#{rails_env} bundle exec sunspot-solr start --port=8983 --data-directory=#{shared_path}/solr/data --pid-dir=#{shared_path}/pids"
  end
  desc "stop solr"
  task :stop, :roles => :app, :except => { :no_release => true } do
    run "cd #{current_path} && RAILS_ENV=#{rails_env} bundle exec sunspot-solr stop --port=8983 --data-directory=#{shared_path}/solr/data --pid-dir=#{shared_path}/pids"
  end
  desc "reindex the whole database"
  task :reindex, :roles => :app do
    stop
    run "rm -rf #{shared_path}/solr/data"
    start
    run "cd #{current_path} && RAILS_ENV=#{rails_env} bundle exec rake sunspot:solr:reindex"
  end
end

after "deploy", "deploy:cleanup", "deploy:migrate_db", 'deploy:setup_solr_data_dir', 'solr:stop', 'solr:reindex', 'solr:start'

I have set the IP for my internal server which is hosting my production app in sunspot.yml:

production:
  solr:
    hostname: [My Server IP]
    port: 8983
    log_level: WARNING
    # read_timeout: 2
    # open_timeout: 0.5

When I try to run cap deploy, I get the following error:

* 2013-04-24 08:28:04 executing `solr:stop'
  * executing "cd /home/username/apps/appname/current && RAILS_ENV=production bundle exec sunspot-solr stop --port=8983 --data-directory=/home/username/apps/appname/shared/solr/data --pid-dir=/home/datacomm/apps/appname/shared/pids"
    servers: ["0.0.0.0"]
    [0.0.0.0] executing command
 ** [out :: 0.0.0.0] java version "1.7.0_15"
 ** [out :: 0.0.0.0] 
 ** [out :: 0.0.0.0] OpenJDK Runtime Environment (IcedTea7 2.3.7) (7u15-2.3.7-0ubuntu1~12.04.1)
 ** [out :: 0.0.0.0] 
 ** [out :: 0.0.0.0] OpenJDK 64-Bit Server VM (build 23.7-b01, mixed mode)
 ** [out :: 0.0.0.0] 
 ** [out :: 0.0.0.0] **No PID file at /home/username/apps/appname/shared/pids/sunspot-solr.pid**
 ** [out :: 0.0.0.0] 
    command finished in 990ms
failed: "rvm_path=$HOME/.rvm/ $HOME/.rvm/bin/rvm-shell '1.9.3' -c 'cd /home/username/apps/appname/current && RAILS_ENV=production bundle exec sunspot-solr stop --port=8983 --data-directory=/home/username/apps/appname/shared/solr/data --pid-dir=/home/username/apps/appname/shared/pids'" on 0.0.0.0

What am I missing? Any help is appreciated.

Answer 1

When run executes the command provided in the solr:stop task, it looks like you're not yet running solr, so their is no pid file for solr present. You're trying to stop a process that doesn't exist.

A Capistrano task is going to throw an exception if the code returned from run is an error code. You can fix this by forcing that command to never return an error.

Make that run command look like this

run("YOUR COMMAND HERE > /dev/null 2>&1 || true")

Now if you call solr:stop when solr isn't running, true will return, letting Capistrano continue on.