terraform/internal/command/graph.go

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

338 lines
11 KiB
Go
Raw Permalink Normal View History

// Copyright IBM Corp. 2014, 2026
// SPDX-License-Identifier: BUSL-1.1
2014-07-01 13:02:13 -04:00
package command
import (
"fmt"
"sort"
2014-07-01 13:02:13 -04:00
"strings"
2015-02-20 01:58:42 -05:00
"github.com/hashicorp/terraform/internal/addrs"
"github.com/hashicorp/terraform/internal/backend/backendrun"
"github.com/hashicorp/terraform/internal/command/arguments"
"github.com/hashicorp/terraform/internal/dag"
core: Functional-style API for terraform.Context Previously terraform.Context was built in an unfortunate way where all of the data was provided up front in terraform.NewContext and then mutated directly by subsequent operations. That made the data flow hard to follow, commonly leading to bugs, and also meant that we were forced to take various actions too early in terraform.NewContext, rather than waiting until a more appropriate time during an operation. This (enormous) commit changes terraform.Context so that its fields are broadly just unchanging data about the execution context (current workspace name, available plugins, etc) whereas the main data Terraform works with arrives via individual method arguments and is returned in return values. Specifically, this means that terraform.Context no longer "has-a" config, state, and "planned changes", instead holding on to those only temporarily during an operation. The caller is responsible for propagating the outcome of one step into the next step so that the data flow between operations is actually visible. However, since that's a change to the main entry points in the "terraform" package, this commit also touches every file in the codebase which interacted with those APIs. Most of the noise here is in updating tests to take the same actions using the new API style, but this also affects the main-code callers in the backends and in the command package. My goal here was to refactor without changing observable behavior, but in practice there are a couple externally-visible behavior variations here that seemed okay in service of the broader goal: - The "terraform graph" command is no longer hooked directly into the core graph builders, because that's no longer part of the public API. However, I did include a couple new Context functions whose contract is to produce a UI-oriented graph, and _for now_ those continue to return the physical graph we use for those operations. There's no exported API for generating the "validate" and "eval" graphs, because neither is particularly interesting in its own right, and so "terraform graph" no longer supports those graph types. - terraform.NewContext no longer has the responsibility for collecting all of the provider schemas up front. Instead, we wait until we need them. However, that means that some of our error messages now have a slightly different shape due to unwinding through a differently-shaped call stack. As of this commit we also end up reloading the schemas multiple times in some cases, which is functionally acceptable but likely represents a performance regression. I intend to rework this to use caching, but I'm saving that for a later commit because this one is big enough already. The proximal reason for this change is to resolve the chicken/egg problem whereby there was previously no single point where we could apply "moved" statements to the previous run state before creating a plan. With this change in place, we can now do that as part of Context.Plan, prior to forking the input state into the three separate state artifacts we use during planning. However, this is at least the third project in a row where the previous API design led to piling more functionality into terraform.NewContext and then working around the incorrect order of operations that produces, so I intend that by paying the cost/risk of this large diff now we can in turn reduce the cost/risk of future projects that relate to our main workflow actions.
2021-08-24 15:06:38 -04:00
"github.com/hashicorp/terraform/internal/plans"
"github.com/hashicorp/terraform/internal/plans/planfile"
"github.com/hashicorp/terraform/internal/terraform"
core: Functional-style API for terraform.Context Previously terraform.Context was built in an unfortunate way where all of the data was provided up front in terraform.NewContext and then mutated directly by subsequent operations. That made the data flow hard to follow, commonly leading to bugs, and also meant that we were forced to take various actions too early in terraform.NewContext, rather than waiting until a more appropriate time during an operation. This (enormous) commit changes terraform.Context so that its fields are broadly just unchanging data about the execution context (current workspace name, available plugins, etc) whereas the main data Terraform works with arrives via individual method arguments and is returned in return values. Specifically, this means that terraform.Context no longer "has-a" config, state, and "planned changes", instead holding on to those only temporarily during an operation. The caller is responsible for propagating the outcome of one step into the next step so that the data flow between operations is actually visible. However, since that's a change to the main entry points in the "terraform" package, this commit also touches every file in the codebase which interacted with those APIs. Most of the noise here is in updating tests to take the same actions using the new API style, but this also affects the main-code callers in the backends and in the command package. My goal here was to refactor without changing observable behavior, but in practice there are a couple externally-visible behavior variations here that seemed okay in service of the broader goal: - The "terraform graph" command is no longer hooked directly into the core graph builders, because that's no longer part of the public API. However, I did include a couple new Context functions whose contract is to produce a UI-oriented graph, and _for now_ those continue to return the physical graph we use for those operations. There's no exported API for generating the "validate" and "eval" graphs, because neither is particularly interesting in its own right, and so "terraform graph" no longer supports those graph types. - terraform.NewContext no longer has the responsibility for collecting all of the provider schemas up front. Instead, we wait until we need them. However, that means that some of our error messages now have a slightly different shape due to unwinding through a differently-shaped call stack. As of this commit we also end up reloading the schemas multiple times in some cases, which is functionally acceptable but likely represents a performance regression. I intend to rework this to use caching, but I'm saving that for a later commit because this one is big enough already. The proximal reason for this change is to resolve the chicken/egg problem whereby there was previously no single point where we could apply "moved" statements to the previous run state before creating a plan. With this change in place, we can now do that as part of Context.Plan, prior to forking the input state into the three separate state artifacts we use during planning. However, this is at least the third project in a row where the previous API design led to piling more functionality into terraform.NewContext and then working around the incorrect order of operations that produces, so I intend that by paying the cost/risk of this large diff now we can in turn reduce the cost/risk of future projects that relate to our main workflow actions.
2021-08-24 15:06:38 -04:00
"github.com/hashicorp/terraform/internal/tfdiags"
2014-07-01 13:02:13 -04:00
)
// GraphCommand is a Command implementation that takes a Terraform
// configuration and outputs the dependency tree in graphical form.
type GraphCommand struct {
Meta
2014-07-01 13:02:13 -04:00
}
func (c *GraphCommand) Run(args []string) int {
var drawCycles bool
2016-12-03 18:00:34 -05:00
var graphTypeStr string
var moduleDepth int
var verbose bool
var planPath string
2014-09-24 20:36:27 -04:00
args = c.Meta.process(args)
cmdFlags := c.Meta.defaultFlagSet("graph")
cmdFlags.BoolVar(&drawCycles, "draw-cycles", false, "draw-cycles")
2016-12-03 18:00:34 -05:00
cmdFlags.StringVar(&graphTypeStr, "type", "", "type")
cmdFlags.IntVar(&moduleDepth, "module-depth", -1, "module-depth")
cmdFlags.BoolVar(&verbose, "verbose", false, "verbose")
cmdFlags.StringVar(&planPath, "plan", "", "plan")
2014-07-01 13:02:13 -04:00
cmdFlags.Usage = func() { c.Ui.Error(c.Help()) }
if err := cmdFlags.Parse(args); err != nil {
c.Ui.Error(fmt.Sprintf("Error parsing command-line flags: %s\n", err.Error()))
2014-07-01 13:02:13 -04:00
return 1
}
configPath, err := ModulePath(cmdFlags.Args())
2017-01-18 23:50:45 -05:00
if err != nil {
c.Ui.Error(err.Error())
return 1
}
// Check for user-supplied plugin path
if c.pluginPath, err = c.loadPluginPath(); err != nil {
c.Ui.Error(fmt.Sprintf("Error loading plugin path: %s", err))
return 1
}
// Try to load plan if path is specified
var planFile *planfile.WrappedPlanFile
if planPath != "" {
planFile, err = c.PlanFile(planPath)
if err != nil {
c.Ui.Error(err.Error())
return 1
}
2017-01-18 23:50:45 -05:00
}
var diags tfdiags.Diagnostics
2017-01-18 23:50:45 -05:00
// Load the backend
PSS: Update how commands access backends, so both `backend` and `state_store` configuration can be used (#37569) * Add a generic method for loading an operations backend in non-init commands * Refactor commands to use new prepareBackend method: group 1 * Refactor commands to use new prepareBackend method: group 2, where config parsing needs to be explicitly added * Refactor commands to use new prepareBackend method: group 3, where we can use already parsed config * Additional, more nested, places where logic for accessing backends needs to be refactored * Remove duplicated comment * Add test coverage of `(m *Meta) prepareBackend()` * Add TODO related to using plans for backend/state_store config in apply commands * Add `testStateStoreMockWithChunkNegotiation` test helper * Add assertions to tests about the backend (remote-state, local, etc) in use within operations backend * Stop prepareBackend taking locks as argument * Code comment in prepareBackend * Replace c.Meta.prepareBackend with c.prepareBackend * Change `c.Meta.loadSingleModule` to `c.loadSingleModule` * Rename (Meta).prepareBackend to (Meta).backend, update godoc comment to make relationship to (Meta).Backend more obvious. * Revert change from config.Module to config.Root.Module * Update `(m *Meta) backend` method to parse config itself, and also to adhere to calling code's viewtype instructions * Update all tests and calling code following previous commit * Change how an operations backend is obtained by autocomplete code * Update autocomplete to return nil if no workspace names are returned from the backend * Add test coverage for autocompleting workspace names when using a pluggable state store * Fix output command: pass view type data to new `backend` method * Fix in plan command: pass correct view type to `backend` method * Fix `providers schema` command to use correct viewtype when preparing a backend
2025-11-03 12:57:20 -05:00
b, backendDiags := c.backend(".", arguments.ViewHuman)
diags = diags.Append(backendDiags)
if backendDiags.HasErrors() {
c.showDiagnostics(diags)
2017-01-18 23:50:45 -05:00
return 1
}
// We require a local backend
local, ok := b.(backendrun.Local)
2017-01-18 23:50:45 -05:00
if !ok {
c.showDiagnostics(diags) // in case of any warnings in here
2017-01-18 23:50:45 -05:00
c.Ui.Error(ErrUnsupportedLocalOp)
return 1
}
backend: Validate remote backend Terraform version When using the enhanced remote backend, a subset of all Terraform operations are supported. Of these, only plan and apply can be executed on the remote infrastructure (e.g. Terraform Cloud). Other operations run locally and use the remote backend for state storage. This causes problems when the local version of Terraform does not match the configured version from the remote workspace. If the two versions are incompatible, an `import` or `state mv` operation can cause the remote workspace to be unusable until a manual fix is applied. To prevent this from happening accidentally, this commit introduces a check that the local Terraform version and the configured remote workspace Terraform version are compatible. This check is skipped for commands which do not write state, and can also be disabled by the use of a new command-line flag, `-ignore-remote-version`. Terraform version compatibility is defined as: - For all releases before 0.14.0, local must exactly equal remote, as two different versions cannot share state; - 0.14.0 to 1.0.x are compatible, as we will not change the state version number until at least Terraform 1.1.0; - Versions after 1.1.0 must have the same major and minor versions, as we will not change the state version number in a patch release. If the two versions are incompatible, a diagnostic is displayed, advising that the error can be suppressed with `-ignore-remote-version`. When this flag is used, the diagnostic is still displayed, but as a warning instead of an error. Commands which will not write state can assert this fact by calling the helper `meta.ignoreRemoteBackendVersionConflict`, which will disable the checks. Those which can write state should instead call the helper `meta.remoteBackendVersionCheck`, which will return diagnostics for display. In addition to these explicit paths for managing the version check, we have an implicit check in the remote backend's state manager initialization method. Both of the above helpers will disable this check. This fallback is in place to ensure that future code paths which access state cannot accidentally skip the remote version check.
2020-11-13 16:43:56 -05:00
// This is a read-only command
c.ignoreRemoteVersionConflict(b)
backend: Validate remote backend Terraform version When using the enhanced remote backend, a subset of all Terraform operations are supported. Of these, only plan and apply can be executed on the remote infrastructure (e.g. Terraform Cloud). Other operations run locally and use the remote backend for state storage. This causes problems when the local version of Terraform does not match the configured version from the remote workspace. If the two versions are incompatible, an `import` or `state mv` operation can cause the remote workspace to be unusable until a manual fix is applied. To prevent this from happening accidentally, this commit introduces a check that the local Terraform version and the configured remote workspace Terraform version are compatible. This check is skipped for commands which do not write state, and can also be disabled by the use of a new command-line flag, `-ignore-remote-version`. Terraform version compatibility is defined as: - For all releases before 0.14.0, local must exactly equal remote, as two different versions cannot share state; - 0.14.0 to 1.0.x are compatible, as we will not change the state version number until at least Terraform 1.1.0; - Versions after 1.1.0 must have the same major and minor versions, as we will not change the state version number in a patch release. If the two versions are incompatible, a diagnostic is displayed, advising that the error can be suppressed with `-ignore-remote-version`. When this flag is used, the diagnostic is still displayed, but as a warning instead of an error. Commands which will not write state can assert this fact by calling the helper `meta.ignoreRemoteBackendVersionConflict`, which will disable the checks. Those which can write state should instead call the helper `meta.remoteBackendVersionCheck`, which will return diagnostics for display. In addition to these explicit paths for managing the version check, we have an implicit check in the remote backend's state manager initialization method. Both of the above helpers will disable this check. This fallback is in place to ensure that future code paths which access state cannot accidentally skip the remote version check.
2020-11-13 16:43:56 -05:00
2017-01-18 23:50:45 -05:00
// Build the operation
opReq := c.Operation(b, arguments.ViewHuman)
opReq.ConfigDir = configPath
opReq.ConfigLoader, err = c.initConfigLoader()
opReq.PlanFile = planFile
opReq.AllowUnsetVariables = true
if err != nil {
diags = diags.Append(err)
c.showDiagnostics(diags)
return 1
}
2017-01-18 23:50:45 -05:00
// Get the context
core: Functional-style API for terraform.Context Previously terraform.Context was built in an unfortunate way where all of the data was provided up front in terraform.NewContext and then mutated directly by subsequent operations. That made the data flow hard to follow, commonly leading to bugs, and also meant that we were forced to take various actions too early in terraform.NewContext, rather than waiting until a more appropriate time during an operation. This (enormous) commit changes terraform.Context so that its fields are broadly just unchanging data about the execution context (current workspace name, available plugins, etc) whereas the main data Terraform works with arrives via individual method arguments and is returned in return values. Specifically, this means that terraform.Context no longer "has-a" config, state, and "planned changes", instead holding on to those only temporarily during an operation. The caller is responsible for propagating the outcome of one step into the next step so that the data flow between operations is actually visible. However, since that's a change to the main entry points in the "terraform" package, this commit also touches every file in the codebase which interacted with those APIs. Most of the noise here is in updating tests to take the same actions using the new API style, but this also affects the main-code callers in the backends and in the command package. My goal here was to refactor without changing observable behavior, but in practice there are a couple externally-visible behavior variations here that seemed okay in service of the broader goal: - The "terraform graph" command is no longer hooked directly into the core graph builders, because that's no longer part of the public API. However, I did include a couple new Context functions whose contract is to produce a UI-oriented graph, and _for now_ those continue to return the physical graph we use for those operations. There's no exported API for generating the "validate" and "eval" graphs, because neither is particularly interesting in its own right, and so "terraform graph" no longer supports those graph types. - terraform.NewContext no longer has the responsibility for collecting all of the provider schemas up front. Instead, we wait until we need them. However, that means that some of our error messages now have a slightly different shape due to unwinding through a differently-shaped call stack. As of this commit we also end up reloading the schemas multiple times in some cases, which is functionally acceptable but likely represents a performance regression. I intend to rework this to use caching, but I'm saving that for a later commit because this one is big enough already. The proximal reason for this change is to resolve the chicken/egg problem whereby there was previously no single point where we could apply "moved" statements to the previous run state before creating a plan. With this change in place, we can now do that as part of Context.Plan, prior to forking the input state into the three separate state artifacts we use during planning. However, this is at least the third project in a row where the previous API design led to piling more functionality into terraform.NewContext and then working around the incorrect order of operations that produces, so I intend that by paying the cost/risk of this large diff now we can in turn reduce the cost/risk of future projects that relate to our main workflow actions.
2021-08-24 15:06:38 -04:00
lr, _, ctxDiags := local.LocalRun(opReq)
diags = diags.Append(ctxDiags)
if ctxDiags.HasErrors() {
c.showDiagnostics(diags)
2014-07-01 13:02:13 -04:00
return 1
}
lr.Core.SetGraphOpts(&terraform.ContextGraphOpts{SkipGraphValidation: drawCycles})
2014-07-01 13:02:13 -04:00
core: Functional-style API for terraform.Context Previously terraform.Context was built in an unfortunate way where all of the data was provided up front in terraform.NewContext and then mutated directly by subsequent operations. That made the data flow hard to follow, commonly leading to bugs, and also meant that we were forced to take various actions too early in terraform.NewContext, rather than waiting until a more appropriate time during an operation. This (enormous) commit changes terraform.Context so that its fields are broadly just unchanging data about the execution context (current workspace name, available plugins, etc) whereas the main data Terraform works with arrives via individual method arguments and is returned in return values. Specifically, this means that terraform.Context no longer "has-a" config, state, and "planned changes", instead holding on to those only temporarily during an operation. The caller is responsible for propagating the outcome of one step into the next step so that the data flow between operations is actually visible. However, since that's a change to the main entry points in the "terraform" package, this commit also touches every file in the codebase which interacted with those APIs. Most of the noise here is in updating tests to take the same actions using the new API style, but this also affects the main-code callers in the backends and in the command package. My goal here was to refactor without changing observable behavior, but in practice there are a couple externally-visible behavior variations here that seemed okay in service of the broader goal: - The "terraform graph" command is no longer hooked directly into the core graph builders, because that's no longer part of the public API. However, I did include a couple new Context functions whose contract is to produce a UI-oriented graph, and _for now_ those continue to return the physical graph we use for those operations. There's no exported API for generating the "validate" and "eval" graphs, because neither is particularly interesting in its own right, and so "terraform graph" no longer supports those graph types. - terraform.NewContext no longer has the responsibility for collecting all of the provider schemas up front. Instead, we wait until we need them. However, that means that some of our error messages now have a slightly different shape due to unwinding through a differently-shaped call stack. As of this commit we also end up reloading the schemas multiple times in some cases, which is functionally acceptable but likely represents a performance regression. I intend to rework this to use caching, but I'm saving that for a later commit because this one is big enough already. The proximal reason for this change is to resolve the chicken/egg problem whereby there was previously no single point where we could apply "moved" statements to the previous run state before creating a plan. With this change in place, we can now do that as part of Context.Plan, prior to forking the input state into the three separate state artifacts we use during planning. However, this is at least the third project in a row where the previous API design led to piling more functionality into terraform.NewContext and then working around the incorrect order of operations that produces, so I intend that by paying the cost/risk of this large diff now we can in turn reduce the cost/risk of future projects that relate to our main workflow actions.
2021-08-24 15:06:38 -04:00
if graphTypeStr == "" {
if planFile == nil {
// Simple resource dependency mode:
// This is based on the plan graph but we then further reduce it down
// to just resource dependency relationships, assuming that in most
// cases the most important thing is what order we'll visit the
// resources in.
fullG, graphDiags := lr.Core.PlanGraphForUI(lr.Config, lr.InputState, plans.NormalMode)
diags = diags.Append(graphDiags)
if graphDiags.HasErrors() {
c.showDiagnostics(diags)
return 1
}
g := fullG.ResourceGraph()
return c.resourceOnlyGraph(g)
} else {
core: Functional-style API for terraform.Context Previously terraform.Context was built in an unfortunate way where all of the data was provided up front in terraform.NewContext and then mutated directly by subsequent operations. That made the data flow hard to follow, commonly leading to bugs, and also meant that we were forced to take various actions too early in terraform.NewContext, rather than waiting until a more appropriate time during an operation. This (enormous) commit changes terraform.Context so that its fields are broadly just unchanging data about the execution context (current workspace name, available plugins, etc) whereas the main data Terraform works with arrives via individual method arguments and is returned in return values. Specifically, this means that terraform.Context no longer "has-a" config, state, and "planned changes", instead holding on to those only temporarily during an operation. The caller is responsible for propagating the outcome of one step into the next step so that the data flow between operations is actually visible. However, since that's a change to the main entry points in the "terraform" package, this commit also touches every file in the codebase which interacted with those APIs. Most of the noise here is in updating tests to take the same actions using the new API style, but this also affects the main-code callers in the backends and in the command package. My goal here was to refactor without changing observable behavior, but in practice there are a couple externally-visible behavior variations here that seemed okay in service of the broader goal: - The "terraform graph" command is no longer hooked directly into the core graph builders, because that's no longer part of the public API. However, I did include a couple new Context functions whose contract is to produce a UI-oriented graph, and _for now_ those continue to return the physical graph we use for those operations. There's no exported API for generating the "validate" and "eval" graphs, because neither is particularly interesting in its own right, and so "terraform graph" no longer supports those graph types. - terraform.NewContext no longer has the responsibility for collecting all of the provider schemas up front. Instead, we wait until we need them. However, that means that some of our error messages now have a slightly different shape due to unwinding through a differently-shaped call stack. As of this commit we also end up reloading the schemas multiple times in some cases, which is functionally acceptable but likely represents a performance regression. I intend to rework this to use caching, but I'm saving that for a later commit because this one is big enough already. The proximal reason for this change is to resolve the chicken/egg problem whereby there was previously no single point where we could apply "moved" statements to the previous run state before creating a plan. With this change in place, we can now do that as part of Context.Plan, prior to forking the input state into the three separate state artifacts we use during planning. However, this is at least the third project in a row where the previous API design led to piling more functionality into terraform.NewContext and then working around the incorrect order of operations that produces, so I intend that by paying the cost/risk of this large diff now we can in turn reduce the cost/risk of future projects that relate to our main workflow actions.
2021-08-24 15:06:38 -04:00
graphTypeStr = "apply"
}
2016-12-03 18:00:34 -05:00
}
core: Functional-style API for terraform.Context Previously terraform.Context was built in an unfortunate way where all of the data was provided up front in terraform.NewContext and then mutated directly by subsequent operations. That made the data flow hard to follow, commonly leading to bugs, and also meant that we were forced to take various actions too early in terraform.NewContext, rather than waiting until a more appropriate time during an operation. This (enormous) commit changes terraform.Context so that its fields are broadly just unchanging data about the execution context (current workspace name, available plugins, etc) whereas the main data Terraform works with arrives via individual method arguments and is returned in return values. Specifically, this means that terraform.Context no longer "has-a" config, state, and "planned changes", instead holding on to those only temporarily during an operation. The caller is responsible for propagating the outcome of one step into the next step so that the data flow between operations is actually visible. However, since that's a change to the main entry points in the "terraform" package, this commit also touches every file in the codebase which interacted with those APIs. Most of the noise here is in updating tests to take the same actions using the new API style, but this also affects the main-code callers in the backends and in the command package. My goal here was to refactor without changing observable behavior, but in practice there are a couple externally-visible behavior variations here that seemed okay in service of the broader goal: - The "terraform graph" command is no longer hooked directly into the core graph builders, because that's no longer part of the public API. However, I did include a couple new Context functions whose contract is to produce a UI-oriented graph, and _for now_ those continue to return the physical graph we use for those operations. There's no exported API for generating the "validate" and "eval" graphs, because neither is particularly interesting in its own right, and so "terraform graph" no longer supports those graph types. - terraform.NewContext no longer has the responsibility for collecting all of the provider schemas up front. Instead, we wait until we need them. However, that means that some of our error messages now have a slightly different shape due to unwinding through a differently-shaped call stack. As of this commit we also end up reloading the schemas multiple times in some cases, which is functionally acceptable but likely represents a performance regression. I intend to rework this to use caching, but I'm saving that for a later commit because this one is big enough already. The proximal reason for this change is to resolve the chicken/egg problem whereby there was previously no single point where we could apply "moved" statements to the previous run state before creating a plan. With this change in place, we can now do that as part of Context.Plan, prior to forking the input state into the three separate state artifacts we use during planning. However, this is at least the third project in a row where the previous API design led to piling more functionality into terraform.NewContext and then working around the incorrect order of operations that produces, so I intend that by paying the cost/risk of this large diff now we can in turn reduce the cost/risk of future projects that relate to our main workflow actions.
2021-08-24 15:06:38 -04:00
var g *terraform.Graph
var graphDiags tfdiags.Diagnostics
switch graphTypeStr {
case "plan":
g, graphDiags = lr.Core.PlanGraphForUI(lr.Config, lr.InputState, plans.NormalMode)
case "plan-refresh-only":
g, graphDiags = lr.Core.PlanGraphForUI(lr.Config, lr.InputState, plans.RefreshOnlyMode)
case "plan-destroy":
g, graphDiags = lr.Core.PlanGraphForUI(lr.Config, lr.InputState, plans.DestroyMode)
case "apply":
plan := lr.Plan
// Historically "terraform graph" would allow the nonsensical request to
// render an apply graph without a plan, so we continue to support that
// here, though perhaps one day this should be an error.
if lr.Plan == nil {
plan = &plans.Plan{
2024-08-22 09:28:00 -04:00
Changes: plans.NewChangesSrc(),
core: Functional-style API for terraform.Context Previously terraform.Context was built in an unfortunate way where all of the data was provided up front in terraform.NewContext and then mutated directly by subsequent operations. That made the data flow hard to follow, commonly leading to bugs, and also meant that we were forced to take various actions too early in terraform.NewContext, rather than waiting until a more appropriate time during an operation. This (enormous) commit changes terraform.Context so that its fields are broadly just unchanging data about the execution context (current workspace name, available plugins, etc) whereas the main data Terraform works with arrives via individual method arguments and is returned in return values. Specifically, this means that terraform.Context no longer "has-a" config, state, and "planned changes", instead holding on to those only temporarily during an operation. The caller is responsible for propagating the outcome of one step into the next step so that the data flow between operations is actually visible. However, since that's a change to the main entry points in the "terraform" package, this commit also touches every file in the codebase which interacted with those APIs. Most of the noise here is in updating tests to take the same actions using the new API style, but this also affects the main-code callers in the backends and in the command package. My goal here was to refactor without changing observable behavior, but in practice there are a couple externally-visible behavior variations here that seemed okay in service of the broader goal: - The "terraform graph" command is no longer hooked directly into the core graph builders, because that's no longer part of the public API. However, I did include a couple new Context functions whose contract is to produce a UI-oriented graph, and _for now_ those continue to return the physical graph we use for those operations. There's no exported API for generating the "validate" and "eval" graphs, because neither is particularly interesting in its own right, and so "terraform graph" no longer supports those graph types. - terraform.NewContext no longer has the responsibility for collecting all of the provider schemas up front. Instead, we wait until we need them. However, that means that some of our error messages now have a slightly different shape due to unwinding through a differently-shaped call stack. As of this commit we also end up reloading the schemas multiple times in some cases, which is functionally acceptable but likely represents a performance regression. I intend to rework this to use caching, but I'm saving that for a later commit because this one is big enough already. The proximal reason for this change is to resolve the chicken/egg problem whereby there was previously no single point where we could apply "moved" statements to the previous run state before creating a plan. With this change in place, we can now do that as part of Context.Plan, prior to forking the input state into the three separate state artifacts we use during planning. However, this is at least the third project in a row where the previous API design led to piling more functionality into terraform.NewContext and then working around the incorrect order of operations that produces, so I intend that by paying the cost/risk of this large diff now we can in turn reduce the cost/risk of future projects that relate to our main workflow actions.
2021-08-24 15:06:38 -04:00
UIMode: plans.NormalMode,
PriorState: lr.InputState,
PrevRunState: lr.InputState,
}
2016-12-03 18:00:34 -05:00
}
core: Functional-style API for terraform.Context Previously terraform.Context was built in an unfortunate way where all of the data was provided up front in terraform.NewContext and then mutated directly by subsequent operations. That made the data flow hard to follow, commonly leading to bugs, and also meant that we were forced to take various actions too early in terraform.NewContext, rather than waiting until a more appropriate time during an operation. This (enormous) commit changes terraform.Context so that its fields are broadly just unchanging data about the execution context (current workspace name, available plugins, etc) whereas the main data Terraform works with arrives via individual method arguments and is returned in return values. Specifically, this means that terraform.Context no longer "has-a" config, state, and "planned changes", instead holding on to those only temporarily during an operation. The caller is responsible for propagating the outcome of one step into the next step so that the data flow between operations is actually visible. However, since that's a change to the main entry points in the "terraform" package, this commit also touches every file in the codebase which interacted with those APIs. Most of the noise here is in updating tests to take the same actions using the new API style, but this also affects the main-code callers in the backends and in the command package. My goal here was to refactor without changing observable behavior, but in practice there are a couple externally-visible behavior variations here that seemed okay in service of the broader goal: - The "terraform graph" command is no longer hooked directly into the core graph builders, because that's no longer part of the public API. However, I did include a couple new Context functions whose contract is to produce a UI-oriented graph, and _for now_ those continue to return the physical graph we use for those operations. There's no exported API for generating the "validate" and "eval" graphs, because neither is particularly interesting in its own right, and so "terraform graph" no longer supports those graph types. - terraform.NewContext no longer has the responsibility for collecting all of the provider schemas up front. Instead, we wait until we need them. However, that means that some of our error messages now have a slightly different shape due to unwinding through a differently-shaped call stack. As of this commit we also end up reloading the schemas multiple times in some cases, which is functionally acceptable but likely represents a performance regression. I intend to rework this to use caching, but I'm saving that for a later commit because this one is big enough already. The proximal reason for this change is to resolve the chicken/egg problem whereby there was previously no single point where we could apply "moved" statements to the previous run state before creating a plan. With this change in place, we can now do that as part of Context.Plan, prior to forking the input state into the three separate state artifacts we use during planning. However, this is at least the third project in a row where the previous API design led to piling more functionality into terraform.NewContext and then working around the incorrect order of operations that produces, so I intend that by paying the cost/risk of this large diff now we can in turn reduce the cost/risk of future projects that relate to our main workflow actions.
2021-08-24 15:06:38 -04:00
g, graphDiags = lr.Core.ApplyGraphForUI(plan, lr.Config)
case "eval", "validate":
// Terraform v0.12 through v1.0 supported both of these, but the
// graph variants for "eval" and "validate" are purely implementation
// details and don't reveal anything (user-model-wise) that you can't
// see in the plan graph.
graphDiags = graphDiags.Append(tfdiags.Sourceless(
tfdiags.Error,
"Graph type no longer available",
fmt.Sprintf("The graph type %q is no longer available. Use -type=plan instead to get a similar result.", graphTypeStr),
))
default:
graphDiags = graphDiags.Append(tfdiags.Sourceless(
tfdiags.Error,
"Unsupported graph type",
`The -type=... argument must be either "plan", "plan-refresh-only", "plan-destroy", or "apply".`,
))
2016-12-03 18:00:34 -05:00
}
terraform: ugly huge change to weave in new HCL2-oriented types Due to how deeply the configuration types go into Terraform Core, there isn't a great way to switch out to HCL2 gradually. As a consequence, this huge commit gets us from the old state to a _compilable_ new state, but does not yet attempt to fix any tests and has a number of known missing parts and bugs. We will continue to iterate on this in forthcoming commits, heading back towards passing tests and making Terraform fully-functional again. The three main goals here are: - Use the configuration models from the "configs" package instead of the older models in the "config" package, which is now deprecated and preserved only to help us write our migration tool. - Do expression inspection and evaluation using the functionality of the new "lang" package, instead of the Interpolator type and related functionality in the main "terraform" package. - Represent addresses of various objects using types in the addrs package, rather than hand-constructed strings. This is not critical to support the above, but was a big help during the implementation of these other points since it made it much more explicit what kind of address is expected in each context. Since our new packages are built to accommodate some future planned features that are not yet implemented (e.g. the "for_each" argument on resources, "count"/"for_each" on modules), and since there's still a fair amount of functionality still using old-style APIs, there is a moderate amount of shimming here to connect new assumptions with old, hopefully in a way that makes it easier to find and eliminate these shims later. I apologize in advance to the person who inevitably just found this huge commit while spelunking through the commit history.
2018-04-30 13:33:53 -04:00
diags = diags.Append(graphDiags)
if graphDiags.HasErrors() {
c.showDiagnostics(diags)
2014-07-01 13:02:13 -04:00
return 1
}
graphStr, err := terraform.GraphDot(g, &dag.DotOpts{
DrawCycles: drawCycles,
MaxDepth: moduleDepth,
Verbose: verbose,
})
if err != nil {
c.Ui.Error(fmt.Sprintf("Error converting graph: %s", err))
return 1
}
if diags.HasErrors() {
// For this command we only show diagnostics if there are errors,
// because printing out naked warnings could upset a naive program
// consuming our dot output.
c.showDiagnostics(diags)
return 1
}
_, err = c.Streams.Stdout.File.WriteString(graphStr)
if err != nil {
c.Ui.Error(fmt.Sprintf("Failed to write graph to stdout: %s", err))
return 1
}
2014-07-01 13:02:13 -04:00
return 0
}
func (c *GraphCommand) resourceOnlyGraph(graph addrs.DirectedGraph[addrs.ConfigResource]) int {
out := c.Streams.Stdout.File
fmt.Fprintln(out, "digraph G {")
// Horizontal presentation is easier to read because our nodes tend
// to be much wider than they are tall. The leftmost nodes in the output
// are those Terraform would visit first.
fmt.Fprintln(out, " rankdir = \"RL\";")
fmt.Fprintln(out, " node [shape = rect, fontname = \"sans-serif\"];")
// To help relate the output back to the configuration it came from,
// and to make the individual node labels more reasonably sized when
// deeply nested inside modules, we'll cluster the nodes together by
// the module they belong to and then show only the local resource
// address in the individual nodes. We'll accomplish that by sorting
// the nodes first by module, so we can then notice the transitions.
allAddrs := graph.AllNodes()
if len(allAddrs) == 0 {
fmt.Fprintln(out, " /* This configuration does not contain any resources. */")
fmt.Fprintln(out, " /* For a more detailed graph, try: terraform graph -type=plan */")
}
addrsOrder := make([]addrs.ConfigResource, 0, len(allAddrs))
for _, addr := range allAddrs {
addrsOrder = append(addrsOrder, addr)
}
sort.Slice(addrsOrder, func(i, j int) bool {
iAddr, jAddr := addrsOrder[i], addrsOrder[j]
iModStr, jModStr := iAddr.Module.String(), jAddr.Module.String()
switch {
case iModStr != jModStr:
return iModStr < jModStr
default:
iRes, jRes := iAddr.Resource, jAddr.Resource
switch {
case iRes.Mode != jRes.Mode:
return iRes.Mode == addrs.DataResourceMode
case iRes.Type != jRes.Type:
return iRes.Type < jRes.Type
default:
return iRes.Name < jRes.Name
}
}
})
currentMod := addrs.RootModule
for _, addr := range addrsOrder {
if !addr.Module.Equal(currentMod) {
// We need a new subgraph, then.
// Experimentally it seems like nested clusters tend to make it
// hard for dot to converge on a good layout, so we'll stick with
// just one level of clusters for now but could revise later based
// on feedback.
if !currentMod.IsRoot() {
fmt.Fprintln(out, " }")
}
currentMod = addr.Module
fmt.Fprintf(out, " subgraph \"cluster_%s\" {\n", currentMod.String())
fmt.Fprintf(out, " label = %q\n", currentMod.String())
fmt.Fprintf(out, " fontname = %q\n", "sans-serif")
}
if currentMod.IsRoot() {
fmt.Fprintf(out, " %q [label=%q];\n", addr.String(), addr.Resource.String())
} else {
fmt.Fprintf(out, " %q [label=%q];\n", addr.String(), addr.Resource.String())
}
}
if !currentMod.IsRoot() {
fmt.Fprintln(out, " }")
}
// Now we'll emit all of the edges.
// We use addrsOrder for both levels to ensure a consistent ordering between
// runs without further sorting, which means we visit more nodes than we
// really need to but this output format is only really useful for relatively
// small graphs anyway, so this should be fine.
for _, sourceAddr := range addrsOrder {
deps := graph.DirectDependenciesOf(sourceAddr)
for _, targetAddr := range addrsOrder {
if !deps.Has(targetAddr) {
continue
}
fmt.Fprintf(out, " %q -> %q;\n", sourceAddr.String(), targetAddr.String())
}
}
fmt.Fprintln(out, "}")
return 0
}
2014-07-01 13:02:13 -04:00
func (c *GraphCommand) Help() string {
helpText := `
Usage: terraform [global options] graph [options]
2014-07-01 13:02:13 -04:00
core: Functional-style API for terraform.Context Previously terraform.Context was built in an unfortunate way where all of the data was provided up front in terraform.NewContext and then mutated directly by subsequent operations. That made the data flow hard to follow, commonly leading to bugs, and also meant that we were forced to take various actions too early in terraform.NewContext, rather than waiting until a more appropriate time during an operation. This (enormous) commit changes terraform.Context so that its fields are broadly just unchanging data about the execution context (current workspace name, available plugins, etc) whereas the main data Terraform works with arrives via individual method arguments and is returned in return values. Specifically, this means that terraform.Context no longer "has-a" config, state, and "planned changes", instead holding on to those only temporarily during an operation. The caller is responsible for propagating the outcome of one step into the next step so that the data flow between operations is actually visible. However, since that's a change to the main entry points in the "terraform" package, this commit also touches every file in the codebase which interacted with those APIs. Most of the noise here is in updating tests to take the same actions using the new API style, but this also affects the main-code callers in the backends and in the command package. My goal here was to refactor without changing observable behavior, but in practice there are a couple externally-visible behavior variations here that seemed okay in service of the broader goal: - The "terraform graph" command is no longer hooked directly into the core graph builders, because that's no longer part of the public API. However, I did include a couple new Context functions whose contract is to produce a UI-oriented graph, and _for now_ those continue to return the physical graph we use for those operations. There's no exported API for generating the "validate" and "eval" graphs, because neither is particularly interesting in its own right, and so "terraform graph" no longer supports those graph types. - terraform.NewContext no longer has the responsibility for collecting all of the provider schemas up front. Instead, we wait until we need them. However, that means that some of our error messages now have a slightly different shape due to unwinding through a differently-shaped call stack. As of this commit we also end up reloading the schemas multiple times in some cases, which is functionally acceptable but likely represents a performance regression. I intend to rework this to use caching, but I'm saving that for a later commit because this one is big enough already. The proximal reason for this change is to resolve the chicken/egg problem whereby there was previously no single point where we could apply "moved" statements to the previous run state before creating a plan. With this change in place, we can now do that as part of Context.Plan, prior to forking the input state into the three separate state artifacts we use during planning. However, this is at least the third project in a row where the previous API design led to piling more functionality into terraform.NewContext and then working around the incorrect order of operations that produces, so I intend that by paying the cost/risk of this large diff now we can in turn reduce the cost/risk of future projects that relate to our main workflow actions.
2021-08-24 15:06:38 -04:00
Produces a representation of the dependency graph between different
objects in the current configuration and state.
2014-07-01 13:02:13 -04:00
By default the graph shows a summary only of the relationships between
resources in the configuration, since those are the main objects that
have side-effects whose ordering is significant. You can generate more
detailed graphs reflecting Terraform's actual evaluation strategy
by specifying the -type=TYPE option to select an operation type.
core: Functional-style API for terraform.Context Previously terraform.Context was built in an unfortunate way where all of the data was provided up front in terraform.NewContext and then mutated directly by subsequent operations. That made the data flow hard to follow, commonly leading to bugs, and also meant that we were forced to take various actions too early in terraform.NewContext, rather than waiting until a more appropriate time during an operation. This (enormous) commit changes terraform.Context so that its fields are broadly just unchanging data about the execution context (current workspace name, available plugins, etc) whereas the main data Terraform works with arrives via individual method arguments and is returned in return values. Specifically, this means that terraform.Context no longer "has-a" config, state, and "planned changes", instead holding on to those only temporarily during an operation. The caller is responsible for propagating the outcome of one step into the next step so that the data flow between operations is actually visible. However, since that's a change to the main entry points in the "terraform" package, this commit also touches every file in the codebase which interacted with those APIs. Most of the noise here is in updating tests to take the same actions using the new API style, but this also affects the main-code callers in the backends and in the command package. My goal here was to refactor without changing observable behavior, but in practice there are a couple externally-visible behavior variations here that seemed okay in service of the broader goal: - The "terraform graph" command is no longer hooked directly into the core graph builders, because that's no longer part of the public API. However, I did include a couple new Context functions whose contract is to produce a UI-oriented graph, and _for now_ those continue to return the physical graph we use for those operations. There's no exported API for generating the "validate" and "eval" graphs, because neither is particularly interesting in its own right, and so "terraform graph" no longer supports those graph types. - terraform.NewContext no longer has the responsibility for collecting all of the provider schemas up front. Instead, we wait until we need them. However, that means that some of our error messages now have a slightly different shape due to unwinding through a differently-shaped call stack. As of this commit we also end up reloading the schemas multiple times in some cases, which is functionally acceptable but likely represents a performance regression. I intend to rework this to use caching, but I'm saving that for a later commit because this one is big enough already. The proximal reason for this change is to resolve the chicken/egg problem whereby there was previously no single point where we could apply "moved" statements to the previous run state before creating a plan. With this change in place, we can now do that as part of Context.Plan, prior to forking the input state into the three separate state artifacts we use during planning. However, this is at least the third project in a row where the previous API design led to piling more functionality into terraform.NewContext and then working around the incorrect order of operations that produces, so I intend that by paying the cost/risk of this large diff now we can in turn reduce the cost/risk of future projects that relate to our main workflow actions.
2021-08-24 15:06:38 -04:00
The graph is presented in the DOT language. The typical program that can
2014-07-11 23:38:03 -04:00
read this format is GraphViz, but many web services are also available
to read this format.
2016-12-03 18:00:34 -05:00
Options:
-plan=tfplan Render graph using the specified plan file instead of the
configuration in the current directory. Implies -type=apply.
-draw-cycles Highlight any cycles in the graph with colored edges.
This helps when diagnosing cycle errors. This option is
supported only when illustrating a real evaluation graph,
selected using the -type=TYPE option.
2014-09-24 20:36:27 -04:00
-type=TYPE Type of operation graph to output. Can be: plan,
plan-refresh-only, plan-destroy, or apply. By default
Terraform just summarizes the relationships between the
resources in your configuration, without any particular
operation in mind. Full operation graphs are more detailed
but therefore often harder to read.
2017-01-26 18:18:42 -05:00
-module-depth=n (deprecated) In prior versions of Terraform, specified the
depth of modules to show in the output.
2014-07-01 13:02:13 -04:00
`
return strings.TrimSpace(helpText)
}
func (c *GraphCommand) Synopsis() string {
return "Generate a Graphviz graph of the steps in an operation"
2014-07-01 13:02:13 -04:00
}