Once we've installed the necessary plugins, we'll do one more walk of
the available plugins and record the SHA256 hashes of all of the plugins
we select in the provider lock file.
The file we write here gets read when we're building ContextOpts to
initialize the main terraform context, so any command that works with
the context will then fail if any of the provider binaries change.
Previously we encouraged users to import a resource and _then_ write the
configuration block for it. This ordering creates lots of risk, since
for various reasons users can end up subsequently running Terraform
without any configuration in place, which then causes Terraform to want
to destroy the resource that was imported.
Now we invert this and require a minimal configuration block be written
first. This helps ensure that the user ends up with a correlated resource
config and state, protecting against any inconsistency caused by typos.
This addresses #11835.
Previously we deferred validation of the resource address on the import
command until we were in the core guts, which caused the error responses
to be rather unhelpful.
By validating these things early we can give better feedback to the user.
This new command prints out the tree of modules annotated with their
associated required providers.
The purpose of this command is to help users answer questions such as
"why is this provider required?", "why is Terraform using an older version
of this provider?", and "what combination of modules is creating an
impossible provider version situation?"
For configurations using many modules this sort of question is likely to
come up a lot once we support versioned providers.
As a bonus use-case, this command also shows explicitly when a provider
configuration is being inherited from a parent module, to help users to
understand where the configuration is coming from for each module when
some child modules provide their own provider configurations.
We're going to use config to determine provider dependencies, so we need
to always provide a config when instantiating a context or we'll end up
loading no providers at all.
We previously had a test for running "terraform import -config=''" to
disable the config entirely, but this test is now removed because it makes
no sense. The actual functionality its testing still remains for now,
but it will be removed in a subsequent commit when we start requiring that
a resource to be imported must already exist in configuration.
Currently this doesn't matter much, but we're about to start checking the
availability of providers early on and so we need to use the correct name
for the mock set of providers we use in command tests, which includes
only a provider named "test".
Without this change, the "push" tests will begin failing once we start
verifying this, since there's no "aws" provider available in the test
context.
This method mirrors that of config.Backend, so we can compare the
configration of a backend read from a config vs that of a backend read
from a state. This will prevent init from reinitializing when using
`-backend-config` options that match the existing state.
The variable validator assumes that any AST node it gets from an
interpolation walk is an indicator of an interpolation. Unfortunately,
back in f223be15 we changed the interpolation walker to emit a LiteralNode
as a way to signal that the result is a literal but not identical to the
input due to escapes.
The existence of this issue suggests a bit of a design smell in that the
interpolation walker interface at first glance appears to skip over all
literals, but it actually emits them in this one situation. In the long
run we should perhaps think about whether the abstraction is right here,
but this is a shallow, tactical change that fixes#13001.
Fixes#12871
We were forgetting to remove the legacy remote state from the actual
state value when migrating. This only causes an issue when saving a plan
since the plan contains the state itself and causes an error where both
a backend + legacy state exist.
If saved plans aren't used this causes no noticable issue.
Due to buggy upgrades already existing in the wild, I also added code to
clear the remote section if it exists in a standard unchanged backend
When migrating from a multi-state backend to a single-state backend, we
have to ensure that our locally configured environment is changed back
to "default", or we won't be able to access the new backend.
This allows a refresh on a non-existent or empty state file. We changed
this in 0.9.0 to error which seemed reasonable but it turns out this
complicates automation that runs refresh since it now needed to
determine if the state file was empty before running.
Its easier to just revert this into a warning with exit code zero.
The reason this changed is because in 0.8.x and earlier, the output
would be simply empty with exit code zero which seemed odd.
Fixes#12749
If we merge in an extra partial config we need to recompute the hash to
compare with the old value to detect that change.
This hash needs to NOT be stored and just used as a temporary. We want
to keep the original hash in the state so that we don't detect a change
from the config (since the config will always be partial).
Fixes#7774
This modifies the `import` command to load configuration files from the
pwd. This also augments the configuration loading section for the CLI to
have a new option (default false, same as old behavior) to
allow directories with no Terraform configurations.
For import, we allow directories with no Terraform configurations so
this option is set to true.
Fixes#7975
This changes the InputMode for the CLI to always be:
InputModeProvider | InputModeVar | InputModeVarUnset
Which means:
* Ask for provider variables
* Ask for user variables _that are not already set_
The change is the latter point. Before, we'd only ask for variables if
zero were given. This forces the user to either have no variables set
via the CLI, env vars, tfvars or ALL variables, but no in between. As
reported in #7975, this isn't expected behavior.
The new change makes is so that unset variables are always asked for.
Users can retain the previous behavior by setting `-input=false`. This
would ensure that variables set by external sources cover all cases.
The handling of remote variables was completely disabled for push.
We still need to fetch variables from atlas for push, because if the
variable is only set remotely the Input walk will still prompt the user
for a value. We add the missing remote variables to the context
to disable input.
We now only handle remote variables as atlas.TFVar and explicitly pass
around that type rather than an `interface{}`.
Shorten the text fixture slightly to make the output a little more
readable on failures.
The strings we have in the variables may contain escaped double-quotes,
which have already been parsed and had the `\`s removed. We need to
re-escape these, but only if we are in the outer string and not inside an
interpolation.
Add tf_vars to the data structures sent in terraform push.
This takes any value of type []interface{} or map[string]interface{} and
marshals it as a string representation of the equivalent HCL. This
prevents ambiguity in atlas between a string that looks like a json
structure, and an actual json structure.
For the time being we will need a way to serialize data as HCL, so the
command package has an internal encodeHCL function to do so. We can
remove this if we get complete package for marshaling HCL.
* core: Add support for marking outputs as sensitive
This commit allows an output to be marked "sensitive", in which case the
value is redacted in the post-refresh and post-apply list of outputs.
For example, the configuration:
```
variable "input" {
default = "Hello world"
}
output "notsensitive" {
value = "${var.input}"
}
output "sensitive" {
sensitive = true
value = "${var.input}"
}
```
Would result in the output:
```
terraform apply
Apply complete! Resources: 0 added, 0 changed, 0 destroyed.
Outputs:
notsensitive = Hello world
sensitive = <sensitive>
```
The `terraform output` command continues to display the value as before.
Limitations: Note that sensitivity is not tracked internally, so if the
output is interpolated in another module into a resource, the value will
be displayed. The value is still present in the state.
Add `-target=resource` flag to core operations, allowing users to
target specific resources in their infrastructure. When `-target` is
used, the operation will only apply to that resource and its
dependencies.
The calculated dependencies are different depending on whether we're
running a normal operation or a `terraform destroy`.
Generally, "dependencies" refers to ancestors: resources falling
_before_ the target in the graph, because their changes are required to
accurately act on the target.
For destroys, "dependencies" are descendents: those resources which fall
_after_ the target. These resources depend on our target, which is going
to be destroyed, so they should also be destroyed.
Radek Simko