We're working on setting up an auto-scalable web service on AWS. One of our requirements is that we can reload the web server configuration via SSH. Usually, this isn't very hard, as it's only a matter of executing
ssh user@webserver "service apache2 reload"
The problem is that, when you auto-scale the web servers, you don't know how many servers are up and, most importantly, what the IP addresses are. So we need a way of getting a list of IP addresses currently spun up in the auto-scaling group.
I've checked out AWS CLI's
describe-auto-scaling-instances command, but that only gives me the instance IDs. AFAIK, there is no way of getting the IP associated with that ID.
I tried searching the AWS API docs for clues, but to no avail.
Is there anyone here who knows of a clean way to do this? The alternative would be that the servers check themselves in at a central server when they boot up, and check themselves out when they spin down. However, this is error-prone, and requires a central server which isn't highly available and frankly a bit of wasteful to use for only this purpose...