This set of changes addresses two bug scenarios:
(1) When an ignored change canceled a resource replacement, any
downstream resources referencing computer attributes on that resource
would get "diffs didn't match" errors. This happened because the
`EvalDiff` implementation was calling `state.MergeDiff(diff)` on the
unfiltered diff. Generally this is what you want, so that downstream
references catch the "incoming" values. When there's a potential for the
diff to change, thought, this results in problems w/ references.
Here we solve this by doing away with the separate `EvalNode` for
`ignore_changes` processing and integrating it into `EvalDiff`. This
allows us to only call `MergeDiff` with the final, filtered diff.
(2) When a resource had an ignored change but was still being replaced
anyways, the diff was being improperly filtered. This would cause
problems during apply when not all attributes were available to perform
the replacement.
We solve that by deferring actual attribute removal until after we've
decided that we do not have to replace the resource.
- Include new option in file provisioner. Now content or source can be
provided. Content will create a temp file and copy there the contents.
- Later that file will be used as source.
- Include test to check that changes are working correctly.
The template resources don't actually need to retain any state, so they
are good candidates to be data sources.
This includes a few tweaks to the acceptance tests -- now configured to
run as unit tests -- since it seems that they have been slightly broken
for a while now. In particular, the "update" cases are no longer tested
because updating is not a meaningful operation for a data source.
As part of evaluating a variable block, there is a pass made on unknown
keys setting them to the config.DefaultVariableValue sentinal value.
Previously this only took into account one level of nesting and assumed
all values were strings.
This commit now traverses the unknown keys via lists and maps and sets
unknown map keys surgically.
Fixes#7241.
Part of the interpolation walk is to detect keys which involve computed
values and therefore cannot be resolved at this time. The interplation
walker keeps sufficient state to be able to populate the ResourceConfig
with a slice of such keys.
Previously they didn't take slice indexes into account, so in the
following case:
```
"services": []interface{}{
map[string]interface{}{
"elb": "___something computed___",
},
map[string]interface{}{
"elb": "___something else computed___",
},
map[string]interface{}{
"elb": "not computed",
},
}
```
Unknown keys would be populated as follows:
```
services.elb
services.elb
```
This is not sufficient information to be useful, as it is impossible to
distinguish which of the `services.elb`s are unknown vs not.
This commit therefore retains the slice indexes as part of the key for
unknown keys - producing for the example above:
```
services.0.elb
services.1.elb
```
The reproduction of issue #7421 involves a list of maps being passed to
a module, where one or more of the maps has a value which is computed
(for example, from another resource). There is a failure at the point of
use (via lookup interpolation) of the computed value of the form:
```
lookup: lookup failed to find 'elb' in:
${lookup(var.services[count.index], "elb")}
```
Where 'elb' is the key of the map.
Previously, any old HTTP error would be treated as the security_group_rule being
deleted. In reality there are only a few cases where this is the right
assumption.
This workaround is originally from 71b30c633f.
According to the commit message from Mitchell Hashimoto:
So I think the AWS API is just broken here. In the case that the state
doesn't update, just assume it did after 5 seconds.
Based on my experience, this AWS API is still broken in the same way.
The timeout was later increased from 5 seconds to 10 seconds in
265cc4fffa.
The timeout (but not the timer) was removed inexplicably in GH-1325.
The symptom is this error from `terraform apply`:
aws_internet_gateway.test: Error waiting for internet gateway (igw-553b4731) to attach: timeout while waiting for state to become '[available]'
followed by all subsequent `terraform apply` commands failing with this error:
aws_internet_gateway.test: Resource.AlreadyAssociated: resource igw-553b4731 is already attached to network vpc-61bc7606
```
make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSRedshiftCluster_importBasic'
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /vendor/)
TF_ACC=1 go test ./builtin/providers/aws -v
-run=TestAccAWSRedshiftCluster_importBasic -timeout 120m
=== RUN TestAccAWSRedshiftCluster_importBasic
--- PASS: TestAccAWSRedshiftCluster_importBasic (623.52s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 623.546s
```
```
% make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSSESReceiptFilter_'
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /vendor/)
TF_ACC=1 go test ./builtin/providers/aws -v
-run=TestAccAWSSESReceiptFilter_ -timeout 120m
=== RUN TestAccAWSSESReceiptFilter_importBasic
--- PASS: TestAccAWSSESReceiptFilter_importBasic (18.18s)
=== RUN TestAccAWSSESReceiptFilter_basic
--- PASS: TestAccAWSSESReceiptFilter_basic (18.42s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 36.633s
```
```
make testacc TEST=./builtin/providers/aws TESTARGS='-run=TestAccAWSSESReceiptRuleSet_'
==> Checking that code complies with gofmt requirements...
go generate $(go list ./... | grep -v /vendor/)
TF_ACC=1 go test ./builtin/providers/aws -v
-run=TestAccAWSSESReceiptRuleSet_ -timeout 120m
=== RUN TestAccAWSSESReceiptRuleSet_importBasic
--- PASS: TestAccAWSSESReceiptRuleSet_importBasic (18.60s)
=== RUN TestAccAWSSESReceiptRuleSet_basic
--- PASS: TestAccAWSSESReceiptRuleSet_basic (26.92s)
PASS
ok github.com/hashicorp/terraform/builtin/providers/aws 45.550s
```
JUst needed some rejigging of the skip_final_snapshot work as that isn't
returned by the API and skipping it means the destroy fails due to
missing final_snapshot_identifier
```
% make testacc TEST=./builtin/providers/aws
% TESTARGS='-run=TestAccAWSDBInstance_' ✹ ✭
==> Checking that code complies with gofmt requirements...
/Users/stacko/Code/go/bin/stringer
go generate $(go list ./... | grep -v /vendor/)
2016/07/07 15:28:31 Generated command/internal_plugin_list.go
TF_ACC=1 go test ./builtin/providers/aws -v -run=TestAccAWSDBInstance_
-timeout 120m
=== RUN TestAccAWSDBInstance_importBasic
--- PASS: TestAccAWSDBInstance_importBasic (588.70s)
=== RUN TestAccAWSDBInstance_basic
--- PASS: TestAccAWSDBInstance_basic (595.71s)
=== RUN TestAccAWSDBInstance_kmsKey
--- PASS: TestAccAWSDBInstance_kmsKey (726.46s)
=== RUN TestAccAWSDBInstance_optionGroup
--- PASS: TestAccAWSDBInstance_optionGroup (681.78s)
=== RUN TestAccAWSDBInstance_iops_update
--- PASS: TestAccAWSDBInstance_iops_update
(590.81s)
```
Please note that I cannot run the enhanced monitoring test in my
environment as I have already got it attached to an IAM role. Running
that test gives me this result:
```
```
When adding multiple notifications from one S3 bucket to one SQS queue, it wasn't immediately intuitive how to do this.
At first I created two `aws_s3_bucket_notification` configs and it seemed to work fine, however the config for one event
will overwrite the other. In order to have multiple events, you can defined the `queue` key twice, or use an array if you're
working with the JSON syntax. I tried to make this more clear in the documentation.
* Fix nested module "unknown variable" during dstry
During a destroy with nested modules, accessing a variable between them
causes an "unknown variable accessed" during destroy.
Passing a literal map to a module looks like this in HCL:
module "foo" {
source = "./foo"
somemap {
somekey = "somevalue"
}
}
The HCL parser always wraps an extra list around the map, so we need to
remove that extra list wrapper when the parameter is indeed of type "map".
Fixes#7140
Radek Simko