Unicorn restarts don't seem to follow unicorn.bogomips.org/SIGNALS.html instructions about capistrano-3-rails-template HOT 2 CLOSED

Hey Victor,

Checkout he chapter on zero downtime deployment, the QUIT signal is sent
one the new processes are ready (implementation is in
https://github.com/TalkingQuickly/capistrano-3-rails-template/blob/master/config/deploy/shared/unicorn.rb.erb
).

This means the old processes are only stopped once the new ones are ready
to take requests so we can deploy new versions of the code seamlessly.

Cheers,

Ben

On Wednesday, 28 October 2015, Víctor M. Valenzuela <
[email protected]> wrote:

Hi, I bought your book and am testing out the provided unicorn_init script.

The issue with it is that stop doesn't always work - only force-stop does.

From the SIGNALS page:

USR2 - reexecute the running binary. A separate QUIT should be sent to the
original process once the child is verified to be up and running

This additional QUIT signal doesn't seem to be sent by the unicorn_init
script:

restart|reload)
sig USR2 && echo reloaded OK && exit 0
echo >&2 "Couldn't reload, starting '$CMD' instead"
run "$CMD"
;;

Should the command consist of something like sig USR2 && oldsig QUIT?

Cheers - Victor

—
Reply to this email directly or view it on GitHub
#41.

Ben Dixon

+44 (0)2 032 894 803
+44 (0)7 762 845 249
http://www.twitter.com/TalkingQuickly
http://www.talkingquickly.co.uk

Dxnx Ltd Trading as Hills Bede
UK Registered Company Number: 6751002

Registered Address:
DXNX LTD
Techhub @ Google Campus
4 - 5 Bonhill Street
London
EC2A 4BX

from capistrano-3-rails-template.

Ahhhh got it, I didn't notice unicorn_init is meant to be used in conjunction with the before_fork after_fork mechanisms.

Some instructions out there such as https://www.digitalocean.com/community/tutorials/how-to-deploy-a-rails-app-with-unicorn-and-nginx-on-ubuntu-14-04 completely omit this (their unicorn.rb has no before_fork/after_fork) :)

Thanks so much!

from capistrano-3-rails-template.