Default is set to 300 seconds, which is AWS default.
This PR fixes a bug (#7898) where instances created within an autoscaling group using the `ec2_asg` module gets immediately terminated because the `health_check_period` is set to 0, which causes the instance to be checked without having the time to actually boot.
Adding `health_check_type` is needed because you may want to check your instance health against an ELB instead of just EC2 default cloudwatch.
Tagging recently created instances can result in
InvalidInstanceID.NotFound errors.
By delaying the tagging until the last part of instance creation,
we should be typically more fortunate (avoiding all such race
conditions might need more work)
- Added a more verbose response
- includes its settable attributes and a list of its instances.
- allows setting of tags, changes upon which mark the task changed
- allow getting of information from asg module, not just setting
- doesn't mark changed if the parameter wasn't specified
- Availability Zones are pulled from the region
For networks that have both a v4 and a v6 subnet, the floating IP plugin
currently has two problems:
* When determining the subnet for the provided `internal_network_name`, it
assumes that the first item in the list of subnets is the one you want.
Instead, it should pick the first v4 subnet.
* When multiple fixed IP's exist for a given port (as is the case in a network
a v4 and a v6 subnet), neutron needs a hint as to which fixed IP to associate
to the floating IP address (the v4 one).
* Fixed error messages to be more descriptive
* Removed direct use of subprocess module and replaced it with calls
to module.run_command
* Changed AZURE_MANAGEMENT_CERT_PATH to be just AZURE_CERT_PATH, which
matches what is expected by the inventory script
fixes error "failed to parse: <attribute 'message' of 'exceptions.BaseException' objects>
TypeError: <attribute 'message' of 'exceptions.BaseException' objects> is not JSON serializable"