Fixes#8541
The Update timeout and delete timeouts were a little short. Bumped them
to 60 mins and 90 mins respectively. I have been on the receiving of the
timeout for the Delete function
* provider/aws: Refresh `aws_elasticsearch_domain` from state when
RecordNotFoundException
Fixes#3967
When an ElasticSearch domain has been deleted outside of Terraform, the
next Terraform operation would return the following:
```
* aws_elasticsearch_domain.curvelogic_es: ResourceNotFoundException:
* Domain not found: curvelogic-es
status code: 409, request id: 6e4b2371-8e1a-11e5-bd07-7741b705d65c
```
We now refresh the resource from state when it is no longer found
* Update resource_aws_elasticsearch_domain.go
`elasticsearch_version` 2.3
Fixes#7836
This will allow ElasticSearch domains to be deployed with version 2.3 of
ElasticSearch
The other slight modifications are to stop dereferencing values before
passing to d.Set in the Read func. It is safer to pass the pointer to
d.Set and allow that to dereference if there is a value
```
% make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSElasticSearchDomain_'
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /terraform/vendor/)
TF_ACC=1 go test ./builtin/providers/aws -v -run=TestAccAWSElasticSearchDomain_ -timeout 120m
=== RUN TestAccAWSElasticSearchDomain_basic
--- PASS: TestAccAWSElasticSearchDomain_basic (1611.74s)
=== RUN TestAccAWSElasticSearchDomain_v23
--- PASS: TestAccAWSElasticSearchDomain_v23 (1898.80s)
=== RUN TestAccAWSElasticSearchDomain_complex
--- PASS: TestAccAWSElasticSearchDomain_complex (1802.44s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 5313.006s
```
Update resource_aws_elasticsearch_domain.go
We create hundreds of AWS Elasticsearch resources over the last few months and we get occasional timeout failures from AWS. This will PR is to increase the timeout once again. I did it before:
https://github.com/hashicorp/terraform/pull/5910/files
But we've seen enough timeouts from AWS on this resource that increasing the timeout seems like the only solution.
Change the `RetryFunc` from a plain `error` return type to a
specialized `RetryError` which must decide whether it is
retryable or not.
Add `RetryableError` / `NonRetryableError` factory functions that
callers are meant to use to build up these errors.
This makes it eminently clear whether or not a given error is
retryable from inside the client code.
Goal here is to _not_ change any behavior, simply reflect the
existing behavior with the new, clearer, API.