Conductor start/stop init-style shell script.

[cwolferh]

Use foreman to start the 3 conductor services. Addresses the
conductor side of
#20
but deltacloud is still a TODO.

[martinpovolny]

checkpid `cat $FOREMAN_PID`

vs.

checkpid `<$FOREMAN_PID`

wins the "useless cat of the day" price although I have heard today that "a command w/o a pipe is not worth running" :-D

or the more readable bracket-%-vim-jumpable, nested-usable variant:

checkpid $(<$FOREMAN_PID)

then the [ -z $variable ] is yet another way to test emptiness introduced into the script, but that has to be unified in some refactoring patch

will put it to a test on some machine next

[martinpovolny]

Fedora 18

• it installs
• it works at the start, but then it stops working

[test@intel-chiefriver-02 pull-request-48]$ bin/conductor stop
Shutting down foreman: OK
[test@intel-chiefriver-02 pull-request-48]$ lsof -i :3033
[test@intel-chiefriver-02 pull-request-48]$ bin/conductor start
Starting foreman: OK
[test@intel-chiefriver-02 pull-request-48]$ lsof -i :3033
[test@intel-chiefriver-02 pull-request-48]$ lsof -i
COMMAND  PID USER   FD   TYPE DEVICE SIZE/OFF NODE NAME
ruby    1311 test    8u  IPv6 145691      0t0  TCP localhost:43162->localhost:postgres (ESTABLISHED)
ruby    1311 test   12u  IPv4 147493      0t0  TCP *:pdb (LISTEN)
ruby    1315 test    8u  IPv6 144916      0t0  TCP localhost:43163->localhost:postgres (ESTABLISHED)
[test@intel-chiefriver-02 pull-request-48]$ bin/conductor stop
Shutting down foreman: FAILED TO SHUTDOWN FOREMAN_PROG
[test@intel-chiefriver-02 pull-request-48]$ bin/conductor start
Starting foreman: OK
[test@intel-chiefriver-02 pull-request-48]$ bin/conductor start
Starting foreman: 
foreman is already running
[test@intel-chiefriver-02 pull-request-48]$ bin/conductor stop
Shutting down foreman: FAILED TO SHUTDOWN FOREMAN_PROG

then it works again and the it stops working again

seems there's some issue when stopping

[cwolferh]

I've modified the stop function in the startup script-- seems to be working better now. Also, I added a restart "shortcut."

[test@localhost ~]$ /home/test/pull-request-48/bin/conductor stop
Shutting down foreman: OK
[test@localhost ~]$ /home/test/pull-request-48/bin/conductor start
Starting foreman: OK
[test@localhost ~]$ /home/test/pull-request-48/bin/conductor start
Starting foreman:
foreman is already running
[test@localhost ~]$ /home/test/pull-request-48/bin/conductor stop
Shutting down foreman: OK
[test@localhost ~]$ /home/test/pull-request-48/bin/conductor stop
Shutting down foreman: /home/test/pull-request-48/conductor/src/tmp/foreman.pid non-existant or process not running
[test@localhost ~]$ /home/test/pull-request-48/bin/conductor restart
Shutting down foreman: /home/test/pull-request-48/conductor/src/tmp/foreman.pid non-existant or process not running
Starting foreman: OK
[test@localhost ~]$ /home/test/pull-request-48/bin/conductor restart
Shutting down foreman: OK
Starting foreman: OK
[test@localhost ~]$ /home/test/pull-request-48/bin/conductor restart
Shutting down foreman: OK
Starting foreman: OK
[test@localhost ~]$ ps -ef | grep -P '(ruby|foreman)'
test     21725     1  0 16:38 pts/2    00:00:00 foreman: master                                                                               \
test     21727 21725 50 16:38 pts/2    00:00:08 /home/test/.rbenv/versions/1.9.3-p374/bin/ruby script/rails s -p 3033
test     21729 21725 46 16:38 pts/2    00:00:07 ruby dbomatic/dbomatic --log log --pid-file tmp -n
test     21731 21725 48 16:38 pts/2    00:00:08 ruby /home/test/pull-request-48/conductor/src/bundle/ruby/1.9.1/bin/rake jobs:work

[martinpovolny]

works as expected