The AWS Go SDK automatically provides a default request retryer with exponential backoff that is invoked via setting `MaxRetries` or leaving it `nil` will default to 3. The terraform-aws-provider `config.Client()` sets `MaxRetries` to 0 unless explicitly configured above 0. Previously, we were not overriding this behavior by setting the configuration and therefore not invoking the default request retryer.
The default retryer already handles HTTP error codes above 500, including S3's InternalError response, so the extraneous handling can be removed. This will also start automatically retrying many additional cases, such as temporary networking issues or other retryable AWS service responses.
Changes:
* s3/backend: Add `max_retries` argument
* s3/backend: Enhance S3 NoSuchBucket error to include additional information
This change enables a few related use cases:
* AWS has partitions outside Commercial, GovCloud (US), and China, which are the only endpoints automatically handled by the AWS Go SDK. DynamoDB locking and credential verification can not currently be enabled in those regions.
* Allows usage of any DynamoDB-compatible API for state locking
* Allows usage of any IAM/STS-compatible API for credential verification
The state manager refactoring in an earlier commit was reflected in the
implementations of these backends, but not in their tests. This gets us
back to a state where the backend tests will compile, and gets _most_ of
them passing again, with a few exceptions that will be addressed in a
subsequent commit.
Due to how often the state and plan types are referenced throughout
Terraform, there isn't a great way to switch them out gradually. As a
consequence, this huge commit gets us from the old world to a _compilable_
new world, but still has a large number of known test failures due to
key functionality being stubbed out.
The stubs here are for anything that interacts with providers, since we
now need to do the follow-up work to similarly replace the old
terraform.ResourceProvider interface with its replacement in the new
"providers" package. That work, along with work to fix the remaining
failing tests, will follow in subsequent commits.
The aim here was to replace all references to terraform.State and its
downstream types with states.State, terraform.Plan with plans.Plan,
state.State with statemgr.State, and switch to the new implementations of
the state and plan file formats. However, due to the number of times those
types are used, this also ended up affecting numerous other parts of core
such as terraform.Hook, the backend.Backend interface, and most of the CLI
commands.
Just as with 5861dbf3fc49b19587a31816eb06f511ab861bb4 before, I apologize
in advance to the person who inevitably just found this huge commit while
spelunking through the commit history.
The new config loader requires some steps to happen in a different
order, particularly in regard to knowing the schema in order to
decode the configuration.
Here we lean directly on the configschema package, rather than
on helper/schema.Backend as before, because it's generally
sufficient for our needs here and this prepares us for the
helper/schema package later moving out into its own repository
to seed a "plugin SDK".
While #16243 added the ability to retry getting a state from S3, Put can
return the same InternalError status. Use the same retry logic when
uploading state to S3.
Internal errors from S3 are usually transient, and can be immediately retried.
Make 2 attempts at retreiving the state object before returning an error.
S3 accepts objects with a leading slash and strips them off. This works
fine except in our workspace hierarchy, which then can no longer find
suffixes matching the full key name.
The s3.Backend was using it's own code for DeleteState, but the dynamo
entries are only handled through the RemoteClient. Have DeleteState use
a RemoteClient for delete.
Since the DynamoDB table used by the S3 backend is no longer only used
for locks, rename it in the config to remove any confusion about it
being lock-specific.
The S3 client can return (nil, nil) when the remote state doesn't exist.
The caused a nil pointer dereference when checking the payload.MD5
against the expected value.
This can happen if the remote state was manually removed, but the digest
entry was left in the DynamoDB table.
Updates to objects in S3 are only eventually consistent. If the
RemoteClient has a DynamoDB table available, use that to store a
checksum of the last written state, so the object can be verified by the
next client to call Get.
Terraform currently doesn't have any sort of user feedback around
RefreshState/Get, so we poll only for a short time before returning an
error.
Prevent extra keys in the s3 envPrefix path from showing up as
listed environments.
Better handle keys containing slashes
Add tests for unexpected keys in s3.
Use the aws provider code to create the clients for the s3 backend, so
that all the behavior matches that of the provider.
Remove the fake creds from the test, as the aws provider will attempt to
validate them.
This adds named state (environment) support to the S3 backend.
A state NAME will prepend the configured s3 key with `env:/NAME/`.
The default state will remain rooted in the bucket for backwards
compatibility.
Locks in DynamoDB use the S3 key as the as the primary key value, so
locking will work as expected for multiple states.
Move the S3 State from a legacy remote state to an official backend.
This increases test coverage, uses a set schema for configuration, and
will allow new backend features to be implemented for the S3 state, e.g.
"environments".
Brian Flad