I ran the first cloud formation and, as described, got the FQDN with working website. From the instructions, I understood that I should simply run the 2nd one and 2 things should happen:
1. it should produce FQDN which is the same as the first script did
2. the original website should stop working
I do get a FQDN which doesn’t work. However, it’s a different one to the one outputted by the first formation. The first one still works BTW.
I didn’t want to look too closely at the files provided to avoid ‘spoilers’, however, I did have a look at the EC2 section and, it seems, that the second script created entirely new webserver (I have 2 instances running).
Now I removed both stacks and deployed only the second one… It did create a web server which doesn’t seem to be responding and I have only one EC2 running. I’m not sure if that’s what I was supposed to have done and I simply misunderstood instructions.. Or do I have only ‘partial deployment’ now and even if I get the net sorted, it’s still not going to work?
Yes, as Ishi says, you are deploying the second CloudFormation on top of the first formation.
The lecture should be updated to clarify the proper steps to take to make the lab work. First, create a new CFN stack using the first YAML file. Once the stack’s status is Create_Complete you need to select that new stack and click the Update button, select "Replace current template", upload second YAML file, then following the same steps to complete the update process. Once the CFN stack’s status is Update_Complete you can test access to the web site. It should be broken.
If you are using the "old" CFN UI you need to select the stack, click the Actions button and select update. This will allow you to update the first stack with the 2nd YAML file to create the issues.
FYI, the CFN templates could use cross-stack references so they could be run as separate stacks. Below is one of many links for more information on cross-stack references:
Is it the person’s inexperience why updating an existing cloudstack should be the correct way to achieve the proper results? Personally, I would rather create a different VPC as not to mess up with an existing "working" one. Assuming that the working one is used a production environment and then updating it with a different stack seems to spell disaster
I had a peek further in the video. In case anyone else has the same problem, you simply have to deploy both with the same name…