I'm confused about what is required to have zero downtime with unicorn, more specifically if it has nothing to do with the preload_app option. I know that if I have preload_app false, I can simply send a HUP signal and unicorn will automatically take into consideration new code, but will it do it in a zero dowtime fashion?
Also, If memory isn't an issue, do I need to use preload_app true at all?
Finally, I see a lot of examples where there is a big before fork block with code about an oldpid. When is this code required?
Thank you
The way I have it set up is to use something like kill -USR2 $(cat /path/to/unicorn.pid)
when restarting the unicorn server, and in my unicorn server config, something like this (based on http://unicorn.bogomips.org/examples/unicorn.conf.rb):
# feel free to point this anywhere accessible on the filesystem
pid "#{shared_path}/pids/unicorn.pid"
before_fork do |server, worker|
# the following is highly recomended for Rails + "preload_app true"
# as there's no need for the master process to hold a connection
defined?(ActiveRecord::Base) and
ActiveRecord::Base.connection.disconnect!
# This allows a new master process to incrementally
# phase out the old master process with SIGTTOU to avoid a
# thundering herd (especially in the "preload_app false" case)
# when doing a transparent upgrade. The last worker spawned
# will then kill off the old master process with a SIGQUIT.
old_pid = "#{server.config[:pid]}.oldbin"
if old_pid != server.pid
begin
sig = (worker.nr + 1) >= server.worker_processes ? :QUIT : :TTOU
Process.kill(sig, File.read(old_pid).to_i)
rescue Errno::ENOENT, Errno::ESRCH
end
end
end
This will start up the new workers and and gradually switch off the old ones.