Because `aws_security_group_rule` resources are an abstraction on top of
Security Groups, they must interact with the AWS Security Group APIs in
a pattern that often results in lots of parallel requests interacting
with the same security group.
We've found that this pattern can trigger race conditions resulting in
inconsistent behavior, including:
* Rules that report as created but don't actually exist on AWS's side
* Rules that show up in AWS but don't register as being created
locally, resulting in follow up attempts to authorize the rule
failing w/ Duplicate errors
Here, we introduce a per-SG mutex that must be held by any security
group before it is allowed to interact with AWS APIs. This protects the
space between `DescribeSecurityGroup` and `Authorize*` / `Revoke*`
calls, ensuring that no other rules interact with the SG during that
span.
The included test exposes the race by applying a security group with
lots of rules, which based on the dependency graph can all be handled in
parallel. This fails most of the time without the new locking behavior.
I've omitted the mutex from `Read`, since it is only called during the
Refresh walk when no changes are being made, meaning a bunch of parallel
`DescribeSecurityGroup` API calls should be consistent in that case.
A "Layer" is a particular service that forms part of the infrastructure for
a set of applications. Some layers are application servers and others are
pure infrastructure, like MySQL servers or load balancers.
Although the AWS API only has one type called "Layer", it actually has
a number of different "soft" types that each have slightly different
validation rules and extra properties that are packed into the Attributes
map.
To make the validation rule differences explicit in Terraform, and to make
the Terraform structure more closely resemble the OpsWorks UI than its
API, we use a separate resource type per layer type, with the common code
factored out into a shared struct type.
"Stack" is the root concept in OpsWorks, and acts as a container for a number
of different "layers" that each provide some service for an application.
A stack isn't very interesting on its own, but it needs to be created before
any layers can be created.
AWS provides three different ways to create AMIs that each have different
inputs, but once they are complete the same management operations apply.
Thus these three resources each have a different "Create" implementation
but then share the same "Read", "Update" and "Delete" implementations.
* master: (720 commits)
Update CHANGELOG.md
Update CHANGELOG.md
dynamodb-local Update AWS config https://github.com/hashicorp/terraform/pull/2825#issuecomment-126353610
Make target_pools optional
Update CHANGELOG.md
code formatting
Update CHANGELOG.md
providers/google: Fix reading account_file path
providers/google: Fix error appending
providers/google: Return if we could parse JSON
providers/google: Change account_file to JSON
providers/google: Default account_file* to empty
providers/google: Add account_file/account_file_contents ConflictsWith
providers/google: Document account_file_contents
providers/google: Use account_file_contents if provided
providers/google: Add account_file_contents to provider
Update CHANGELOG.md
Update CHANGELOG.md
dynamodb-local Use ` instead of : to refer region to keep the consistency with the provider docs
dynamodb-local Update aws provider docs to include the `dynamodb_endpoint` argument
...
stack72