Improve AWS installation documentation
Some of the steps in the AWS installation documentation are unclear or out of order.
-
The pg_trgm
step assumes that the user has a machine that can connect to the database, but there's no such machine at that point -
The "After deployment" steps assume that the user can shell to the server(s), but they can't (there's no public IP) and anyway the servers are being launched by an autoscaling group (so editing gitlab.rb is pointless) -
Load Balancer section needs better instructions on when/how to set up the Target Group. Step 8 of the instructions refers to an "existing target group", but nowhere above that line does it take the user through that process. That needs to be done on the EC2 Dashboard in the Target Group menu prior to setting up the Load Balancer. -
The command "gitlab-ctl reconfigure" will hang indefinitely when the gitlab_rails database configurations are incorrect. I've found that interrupting the nonresponsive reconfigure task has rendered my gitlab instance inoperable and in an irrecoverable state. I was unable to find any adequate way to remedy this situation, instead having to terminate my newly provisioned EC2 instance and start by configuring a new clone of your EE starter image again. I had to repeat this process 4-5 times as I whittled down the possible causes for the reconfiguration to fail, which was a very painful process. -
Needs instructions for configuring SES to send email
Edited by Francis Potter (personal)