Using Rails 3.2 and Thinking Sphinx 3.0.6. When I checked the processes running on my production environment, I got this:
deployer@vps:~$ ps aux | grep search
deployer 401 0.0 0.0 102340 1588 ? S 10:38 0:00 searchd --pidfile --config /home/deployer/abc/releases/1/config/production.sphinx.conf
deployer 402 47.0 0.7 248984 29944 ? Sl 10:38 0:12 searchd --pidfile --config /home/deployer/abc/releases/1/config/production.sphinx.conf
There are 2 processes in Sphinx. Is this behavior right?
Everytime when I run ts:stop, it hangs, and I have to kill one of the processes manually to stop Sphinx. I'm not sure if this is caused by the two processes.
Update 1
My thinking_sphinx.yml:
production:
pid_file: /home/deployer/abc/shared/pids/searchd.pid
indices_location: /home/deployer/abc/shared/db/sphinx
searchd_file_path: /home/deployer/abc/shared/db/sphinx
config_file: /home/deployer/abc/shared/config/production.sphinx.conf
mysql41: 9312
mem_limit: 512M
Two processes is normal - TS v3 uses Sphinx's threaded workers (so, you're seeing the master process and at least one child).
As for the ts:stop issue - I presume this is via Capistrano? It's probably because the configuration file doesn't exist for the current deployed release. You should probably change the file location (using the configuration_file setting) to somewhere in the shared directory.
This is covered in the deployment documentation (but was only added about a month ago).
As for the yml settings - you can remove searchd_file_path, as it's been replaced by indices_location in TS v3, and config_file should be configuration_file.