I need to change the detach version of the task to use Kernel.exec, I think - that way, the process changes to searchd, instead of searchd being run within rake.
If you want to have a stab at patching, you're certainly welcome to give it a shot :) -- Pat On 18/11/2011, at 6:48 PM, Andrew Stewart wrote: > Hi Pat, > > On 18 Nov 2011, at 12:13, Pat Allan wrote: >> What happens if you run searchd manually instead of via the rake task? >> >> searchd --config path/to/production.sphinx.conf --nodetach > > I patched my Upstart configuration to run searchd directly instead of via > Rake...it worked perfectly and barely took any CPU at all. > > (I also realised that my pid_file location in sphinx.yml was unwriteable; > once I'd fixed that, Sphinx was able to run. Maybe that was pinning the CPU?) > > When I revert to the Rake method the CPU starts at >95% but declines over > several minutes. After 5 minutes my process list looks like: > > $ ps wwaux | grep sphinx > rails 31907 0.0 0.1 35636 1240 ? Ss 11:39 0:00 su - rails > -c cd /var/www/apps/sparkle/releases/20111118082310; export PORT=5100; export > RAILS_ENV=production; bundle exec rake thinking_sphinx:start NODETACH=true > >> /var/www/apps/sparkle/shared/log/thinking_sphinx-1.log 2>&1 > rails 31912 0.0 0.2 19780 2312 ? S 11:39 0:00 -su -c cd > /var/www/apps/sparkle/releases/20111118082310; export PORT=5100; export > RAILS_ENV=production; bundle exec rake thinking_sphinx:start NODETACH=true > >> /var/www/apps/sparkle/shared/log/thinking_sphinx-1.log 2>&1 > rails 32048 4.8 8.1 285084 83512 ? Sl 11:39 0:16 ruby > /var/www/apps/sparkle/shared/bundle/ruby/1.9.1/bin/rake thinking_sphinx:start > NODETACH=true > rails 32136 0.0 0.0 4144 564 ? S 11:39 0:00 sh -c > /usr/local/bin/searchd --pidfile --config > "/var/www/apps/sparkle/releases/20111118082310/config/production.sphinx.conf" > --nodetach > rails 32139 0.0 0.4 37828 5040 ? S 11:39 0:00 > /usr/local/bin/searchd --pidfile --config > /var/www/apps/sparkle/releases/20111118082310/config/production.sphinx.conf > --nodetach > > So the CPU is down to 4.8%. Much better than 95% but not as good as > bypassing Rake altogether. Speaking of Rake, I wonder why it is still > hanging around? > > Cheers, > Andrew > > -- > You received this message because you are subscribed to the Google Groups > "Thinking Sphinx" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to > [email protected]. > For more options, visit this group at > http://groups.google.com/group/thinking-sphinx?hl=en. > -- You received this message because you are subscribed to the Google Groups "Thinking Sphinx" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/thinking-sphinx?hl=en.
