StackOverflow2013

Note that there are some explanatory texts on larger screens.

plurals

POBest practice for Chef, ec2, knife and reducing the time to boot and ami in the event of a failover
text
Body
copied!<p>I an in the process of placing all of my servers under the control of chef for scm using chef server. It works great....I love it. I think I am starting to get the zen but I stil need a better understanding.</p> <p>Prior, on ec2 i built ami's and I was always rebuilding ami's but at least they boot fast.</p> <p>With chef, i gather, is to always start with a barebones instance and use knife to build your ami.</p> <p>It takes about 20 minutes to launch an ami of which installing chef is about 5-6 minutes.</p> <p>Issue I have is that I am running 4 servers, two redis, master and slave and two nginx servers under load balancing.</p> <p>On a remote machine I have python scripts that are always checking if a a redis master or slave is down and for the load balancer if the nginx servers if they are healthy. </p> <p>Now, under chef, if a server goes down I have to wait an extra 15+ minutes where under pre-built ami's the length of time was the time to boot up a new instance.</p> <p>Questions I have are the following:</p> <p>1) Is there a python api to interact with knife? In python I used boto to boot an ami. Now, I will have to use a popen to execute knife. My code depends on knowing when the ami is up and running to continue further processing. </p> <p>2) Is it best practice to have an ami where chef client is pre installed? That would shave about 6 minites off.</p> <p>3) finally, what else am I missing? Any advice would be great. </p>

Querying!

Guidance

An individual column

Larger individual text columns get their own page to allow for proper reading.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload