Investigation completed.
It looks like the CloudFront Distributions that s2Member configured, were somehow accessed by another software application, which is using a different version of the CloudFront API. This seems to cause some problems on the CloudFront side of things, due to some recent changes/updates from Amazon.
Amazon just recently released a whole new API for their services. Some software applications (like s2Member) are still using the previous version, and some are now using the new version. If two different software applications attempt to configure CloudFront distributions, but in different ways (i.e. with two different APIs), it forces future API requests to occur with the latest version; else they will fail.
I believe this is the underlying cause of the problems that you’re experiencing. s2Member is NOT using the latest version of the Amazon API. It was only recently released, and while work is underway to keep things updated, we’re not there yet. s2Member uses API version: 2010-11-01 (the previous version).
The newest version is: 2012-05-05.
Here’s what I suggest.
In most scenarios, ONLY s2Member is going to access your CloudFront Distribution configurations. Therefore, I would simply start over from scratch, and let s2Member re-configure things for you.
1. Log into your Amazon CloudFront console at Amazon.com.
2. Manually disable (and then delete), both of your existing s2/CloudFront Distributions.
3. Then come back to your installation of s2Member, and check the box to re-configure CloudFront.
This will give you a fresh start, and s2Member’s integration should start working for you again, as expected.
Preventing this from happening again.
* Please do NOT configure another software application, running on a different API version, to communicate with these CloudFront Distributions. This seems to cause a conflict on the Amazon side of things, and future API calls may fail. Hopefully they’ll get this fixed asap. I’m reporting this now.