[FEATURE] ec2 to ec2_instance

[FredButler]

Move from ec2 to amazon.aws.ec2_instance module. volumes structure changed for ec2_instance; maintain old syntax in cluster_defs. Spot instances no longer supported

Chrony changes showing although already merged

[dseeley]

Looks familiar (c.f. https://github.com/dseeley/clusterverse/commit/e832c5f274d04d2e21559be4debc27dbe183fd8e)!!

There are a few other changes & fixes in there that might be useful, e.g.

• amazon.aws.ec2_instance doesn't return the tagged_instances structure if the instances already exist, so the playbook would not be idempotent (& I think this kills one or two of the redeploy schemes).
• You need Ansible core >=2.12.4 (or pypi v5.6.0) now, so the preflight assertions should change (they actually get a bit simpler).
• I found that the route53 rescue mechanism (create_dns_a.yml) wasn't working, as the module_args were not present anymore, hence these values are acquired from elsewhere.

[FredButler]

Looks familiar (c.f. dseeley@e832c5f)!!

There are a few other changes & fixes in there that might be useful, e.g.

• amazon.aws.ec2_instance doesn't return the tagged_instances structure if the instances already exist, so the playbook would not be idempotent (& I think this kills one or two of the redeploy schemes).
• You need Ansible core >=2.12.4 (or pypi v5.6.0) now, so the preflight assertions should change (they actually get a bit simpler).
• I found that the route53 rescue mechanism (create_dns_a.yml) wasn't working, as the module_args were not present anymore, hence these values are acquired from elsewhere.

Thanks for having a look @dseeley, I did find your PR albeit a good time after struggling without it, defiantly helped. create_dns_a.yml was an issue. I'll look into this.

I'd need to test the redeploy schemes, if we still use those that is.

Thanks for the hint of the preflight assertions also.