<p dir="ltr"><br>
On 10 Apr 2016 19:03, "Kasper Grubbe" <<a href="mailto:kawsper@gmail.com">kawsper@gmail.com</a>> wrote:<br>
><br>
> I don't understand the Ali G joke/reference/etc. tagline.</p>
<p dir="ltr">Sorry, my sense of humour is terrible.</p>
<p dir="ltr">><br>
> It could be interesting to dive a bit further into using it with Unicorn, and go a bit more in depth with deployments. I believe you would be able to achieve zero-downtime deploys when reloading new code by issuing `systemctl reload app`.<br>
></p>
<p dir="ltr">Indeed, you'd have to configure the init script to send a specific signal to the server on a reload. I'm sure it is just a configuration option.</p>
<p dir="ltr">Another strategy, which I've had great success in using, is to bring up a brand new server cluster on every deploy and then move traffic over to it without interruptions.</p>
<p dir="ltr">All the best.<br>
</p>