command: Unmanaged providers
This adds supports for "unmanaged" providers, or providers with process
lifecycles not controlled by Terraform. These providers are assumed to
be started before Terraform is launched, and are assumed to shut
themselves down after Terraform has finished running.
To do this, we must update the go-plugin dependency to v1.3.0, which
added support for the "test mode" plugin serving that powers all this.
As a side-effect of not needing to manage the process lifecycle anymore,
Terraform also no longer needs to worry about the provider's binary, as
it won't be used for anything anymore. Because of this, we can disable
the init behavior that concerns itself with downloading that provider's
binary, checking its version, and otherwise managing the binary.
This is all managed on a per-provider basis, so managed providers that
Terraform downloads, starts, and stops can be used in the same commands
as unmanaged providers. The TF_REATTACH_PROVIDERS environment variable
is added, and is a JSON encoding of the provider's address to the
information we need to connect to it.
This change enables two benefits: first, delve and other debuggers can
now be attached to provider server processes, and Terraform can connect.
This allows for attaching debuggers to provider processes, which before
was difficult to impossible. Second, it allows the SDK test framework to
host the provider in the same process as the test driver, while running
a production Terraform binary against the provider. This allows for Go's
built-in race detector and test coverage tooling to work as expected in
provider tests.
Unmanaged providers are expected to work in the exact same way as
managed providers, with one caveat: Terraform kills provider processes
and restarts them once per graph walk, meaning multiple times during
most Terraform CLI commands. As unmanaged providers can't be killed by
Terraform, and have no visibility into graph walks, unmanaged providers
are likely to have differences in how their global mutable state behaves
when compared to managed providers. Namely, unmanaged providers are
likely to retain global state when managed providers would have reset
it. Developers relying on global state should be aware of this.
2020-05-27 02:48:57 +02:00
|
|
|
package e2etest
|
|
|
|
|
|
|
|
import (
|
|
|
|
"context"
|
|
|
|
"encoding/json"
|
|
|
|
"io/ioutil"
|
|
|
|
"path/filepath"
|
|
|
|
"strings"
|
2020-09-30 20:28:02 +02:00
|
|
|
"sync"
|
command: Unmanaged providers
This adds supports for "unmanaged" providers, or providers with process
lifecycles not controlled by Terraform. These providers are assumed to
be started before Terraform is launched, and are assumed to shut
themselves down after Terraform has finished running.
To do this, we must update the go-plugin dependency to v1.3.0, which
added support for the "test mode" plugin serving that powers all this.
As a side-effect of not needing to manage the process lifecycle anymore,
Terraform also no longer needs to worry about the provider's binary, as
it won't be used for anything anymore. Because of this, we can disable
the init behavior that concerns itself with downloading that provider's
binary, checking its version, and otherwise managing the binary.
This is all managed on a per-provider basis, so managed providers that
Terraform downloads, starts, and stops can be used in the same commands
as unmanaged providers. The TF_REATTACH_PROVIDERS environment variable
is added, and is a JSON encoding of the provider's address to the
information we need to connect to it.
This change enables two benefits: first, delve and other debuggers can
now be attached to provider server processes, and Terraform can connect.
This allows for attaching debuggers to provider processes, which before
was difficult to impossible. Second, it allows the SDK test framework to
host the provider in the same process as the test driver, while running
a production Terraform binary against the provider. This allows for Go's
built-in race detector and test coverage tooling to work as expected in
provider tests.
Unmanaged providers are expected to work in the exact same way as
managed providers, with one caveat: Terraform kills provider processes
and restarts them once per graph walk, meaning multiple times during
most Terraform CLI commands. As unmanaged providers can't be killed by
Terraform, and have no visibility into graph walks, unmanaged providers
are likely to have differences in how their global mutable state behaves
when compared to managed providers. Namely, unmanaged providers are
likely to retain global state when managed providers would have reset
it. Developers relying on global state should be aware of this.
2020-05-27 02:48:57 +02:00
|
|
|
"testing"
|
|
|
|
|
|
|
|
"github.com/hashicorp/go-hclog"
|
|
|
|
"github.com/hashicorp/go-plugin"
|
|
|
|
"github.com/hashicorp/terraform/builtin/providers/test"
|
|
|
|
"github.com/hashicorp/terraform/e2e"
|
|
|
|
grpcplugin "github.com/hashicorp/terraform/helper/plugin"
|
|
|
|
proto "github.com/hashicorp/terraform/internal/tfplugin5"
|
|
|
|
tfplugin "github.com/hashicorp/terraform/plugin"
|
|
|
|
)
|
|
|
|
|
|
|
|
// The tests in this file are for the "unmanaged provider workflow", which
|
|
|
|
// includes variants of the following sequence, with different details:
|
|
|
|
// terraform init
|
|
|
|
// terraform plan
|
|
|
|
// terraform apply
|
|
|
|
//
|
|
|
|
// These tests are run against an in-process server, and checked to make sure
|
|
|
|
// they're not trying to control the lifecycle of the binary. They are not
|
|
|
|
// checked for correctness of the operations themselves.
|
|
|
|
|
|
|
|
type reattachConfig struct {
|
|
|
|
Protocol string
|
|
|
|
Pid int
|
|
|
|
Test bool
|
|
|
|
Addr reattachConfigAddr
|
|
|
|
}
|
|
|
|
|
|
|
|
type reattachConfigAddr struct {
|
|
|
|
Network string
|
|
|
|
String string
|
|
|
|
}
|
|
|
|
|
|
|
|
type providerServer struct {
|
2020-09-30 20:28:02 +02:00
|
|
|
sync.Mutex
|
command: Unmanaged providers
This adds supports for "unmanaged" providers, or providers with process
lifecycles not controlled by Terraform. These providers are assumed to
be started before Terraform is launched, and are assumed to shut
themselves down after Terraform has finished running.
To do this, we must update the go-plugin dependency to v1.3.0, which
added support for the "test mode" plugin serving that powers all this.
As a side-effect of not needing to manage the process lifecycle anymore,
Terraform also no longer needs to worry about the provider's binary, as
it won't be used for anything anymore. Because of this, we can disable
the init behavior that concerns itself with downloading that provider's
binary, checking its version, and otherwise managing the binary.
This is all managed on a per-provider basis, so managed providers that
Terraform downloads, starts, and stops can be used in the same commands
as unmanaged providers. The TF_REATTACH_PROVIDERS environment variable
is added, and is a JSON encoding of the provider's address to the
information we need to connect to it.
This change enables two benefits: first, delve and other debuggers can
now be attached to provider server processes, and Terraform can connect.
This allows for attaching debuggers to provider processes, which before
was difficult to impossible. Second, it allows the SDK test framework to
host the provider in the same process as the test driver, while running
a production Terraform binary against the provider. This allows for Go's
built-in race detector and test coverage tooling to work as expected in
provider tests.
Unmanaged providers are expected to work in the exact same way as
managed providers, with one caveat: Terraform kills provider processes
and restarts them once per graph walk, meaning multiple times during
most Terraform CLI commands. As unmanaged providers can't be killed by
Terraform, and have no visibility into graph walks, unmanaged providers
are likely to have differences in how their global mutable state behaves
when compared to managed providers. Namely, unmanaged providers are
likely to retain global state when managed providers would have reset
it. Developers relying on global state should be aware of this.
2020-05-27 02:48:57 +02:00
|
|
|
*grpcplugin.GRPCProviderServer
|
|
|
|
planResourceChangeCalled bool
|
|
|
|
applyResourceChangeCalled bool
|
|
|
|
}
|
|
|
|
|
|
|
|
func (p *providerServer) PlanResourceChange(ctx context.Context, req *proto.PlanResourceChange_Request) (*proto.PlanResourceChange_Response, error) {
|
2020-09-30 20:28:02 +02:00
|
|
|
p.Lock()
|
|
|
|
defer p.Unlock()
|
|
|
|
|
command: Unmanaged providers
This adds supports for "unmanaged" providers, or providers with process
lifecycles not controlled by Terraform. These providers are assumed to
be started before Terraform is launched, and are assumed to shut
themselves down after Terraform has finished running.
To do this, we must update the go-plugin dependency to v1.3.0, which
added support for the "test mode" plugin serving that powers all this.
As a side-effect of not needing to manage the process lifecycle anymore,
Terraform also no longer needs to worry about the provider's binary, as
it won't be used for anything anymore. Because of this, we can disable
the init behavior that concerns itself with downloading that provider's
binary, checking its version, and otherwise managing the binary.
This is all managed on a per-provider basis, so managed providers that
Terraform downloads, starts, and stops can be used in the same commands
as unmanaged providers. The TF_REATTACH_PROVIDERS environment variable
is added, and is a JSON encoding of the provider's address to the
information we need to connect to it.
This change enables two benefits: first, delve and other debuggers can
now be attached to provider server processes, and Terraform can connect.
This allows for attaching debuggers to provider processes, which before
was difficult to impossible. Second, it allows the SDK test framework to
host the provider in the same process as the test driver, while running
a production Terraform binary against the provider. This allows for Go's
built-in race detector and test coverage tooling to work as expected in
provider tests.
Unmanaged providers are expected to work in the exact same way as
managed providers, with one caveat: Terraform kills provider processes
and restarts them once per graph walk, meaning multiple times during
most Terraform CLI commands. As unmanaged providers can't be killed by
Terraform, and have no visibility into graph walks, unmanaged providers
are likely to have differences in how their global mutable state behaves
when compared to managed providers. Namely, unmanaged providers are
likely to retain global state when managed providers would have reset
it. Developers relying on global state should be aware of this.
2020-05-27 02:48:57 +02:00
|
|
|
p.planResourceChangeCalled = true
|
|
|
|
return p.GRPCProviderServer.PlanResourceChange(ctx, req)
|
|
|
|
}
|
|
|
|
|
|
|
|
func (p *providerServer) ApplyResourceChange(ctx context.Context, req *proto.ApplyResourceChange_Request) (*proto.ApplyResourceChange_Response, error) {
|
2020-09-30 20:28:02 +02:00
|
|
|
p.Lock()
|
|
|
|
defer p.Unlock()
|
|
|
|
|
command: Unmanaged providers
This adds supports for "unmanaged" providers, or providers with process
lifecycles not controlled by Terraform. These providers are assumed to
be started before Terraform is launched, and are assumed to shut
themselves down after Terraform has finished running.
To do this, we must update the go-plugin dependency to v1.3.0, which
added support for the "test mode" plugin serving that powers all this.
As a side-effect of not needing to manage the process lifecycle anymore,
Terraform also no longer needs to worry about the provider's binary, as
it won't be used for anything anymore. Because of this, we can disable
the init behavior that concerns itself with downloading that provider's
binary, checking its version, and otherwise managing the binary.
This is all managed on a per-provider basis, so managed providers that
Terraform downloads, starts, and stops can be used in the same commands
as unmanaged providers. The TF_REATTACH_PROVIDERS environment variable
is added, and is a JSON encoding of the provider's address to the
information we need to connect to it.
This change enables two benefits: first, delve and other debuggers can
now be attached to provider server processes, and Terraform can connect.
This allows for attaching debuggers to provider processes, which before
was difficult to impossible. Second, it allows the SDK test framework to
host the provider in the same process as the test driver, while running
a production Terraform binary against the provider. This allows for Go's
built-in race detector and test coverage tooling to work as expected in
provider tests.
Unmanaged providers are expected to work in the exact same way as
managed providers, with one caveat: Terraform kills provider processes
and restarts them once per graph walk, meaning multiple times during
most Terraform CLI commands. As unmanaged providers can't be killed by
Terraform, and have no visibility into graph walks, unmanaged providers
are likely to have differences in how their global mutable state behaves
when compared to managed providers. Namely, unmanaged providers are
likely to retain global state when managed providers would have reset
it. Developers relying on global state should be aware of this.
2020-05-27 02:48:57 +02:00
|
|
|
p.applyResourceChangeCalled = true
|
|
|
|
return p.GRPCProviderServer.ApplyResourceChange(ctx, req)
|
|
|
|
}
|
|
|
|
|
2020-09-30 20:28:02 +02:00
|
|
|
func (p *providerServer) PlanResourceChangeCalled() bool {
|
|
|
|
p.Lock()
|
|
|
|
defer p.Unlock()
|
|
|
|
|
|
|
|
return p.planResourceChangeCalled
|
|
|
|
}
|
|
|
|
func (p *providerServer) ResetPlanResourceChangeCalled() {
|
|
|
|
p.Lock()
|
|
|
|
defer p.Unlock()
|
|
|
|
|
|
|
|
p.planResourceChangeCalled = false
|
|
|
|
}
|
|
|
|
|
|
|
|
func (p *providerServer) ApplyResourceChangeCalled() bool {
|
|
|
|
p.Lock()
|
|
|
|
defer p.Unlock()
|
|
|
|
|
|
|
|
return p.applyResourceChangeCalled
|
|
|
|
}
|
|
|
|
func (p *providerServer) ResetApplyResourceChangeCalled() {
|
|
|
|
p.Lock()
|
|
|
|
defer p.Unlock()
|
|
|
|
|
|
|
|
p.applyResourceChangeCalled = false
|
|
|
|
}
|
|
|
|
|
command: Unmanaged providers
This adds supports for "unmanaged" providers, or providers with process
lifecycles not controlled by Terraform. These providers are assumed to
be started before Terraform is launched, and are assumed to shut
themselves down after Terraform has finished running.
To do this, we must update the go-plugin dependency to v1.3.0, which
added support for the "test mode" plugin serving that powers all this.
As a side-effect of not needing to manage the process lifecycle anymore,
Terraform also no longer needs to worry about the provider's binary, as
it won't be used for anything anymore. Because of this, we can disable
the init behavior that concerns itself with downloading that provider's
binary, checking its version, and otherwise managing the binary.
This is all managed on a per-provider basis, so managed providers that
Terraform downloads, starts, and stops can be used in the same commands
as unmanaged providers. The TF_REATTACH_PROVIDERS environment variable
is added, and is a JSON encoding of the provider's address to the
information we need to connect to it.
This change enables two benefits: first, delve and other debuggers can
now be attached to provider server processes, and Terraform can connect.
This allows for attaching debuggers to provider processes, which before
was difficult to impossible. Second, it allows the SDK test framework to
host the provider in the same process as the test driver, while running
a production Terraform binary against the provider. This allows for Go's
built-in race detector and test coverage tooling to work as expected in
provider tests.
Unmanaged providers are expected to work in the exact same way as
managed providers, with one caveat: Terraform kills provider processes
and restarts them once per graph walk, meaning multiple times during
most Terraform CLI commands. As unmanaged providers can't be killed by
Terraform, and have no visibility into graph walks, unmanaged providers
are likely to have differences in how their global mutable state behaves
when compared to managed providers. Namely, unmanaged providers are
likely to retain global state when managed providers would have reset
it. Developers relying on global state should be aware of this.
2020-05-27 02:48:57 +02:00
|
|
|
func TestUnmanagedSeparatePlan(t *testing.T) {
|
|
|
|
t.Parallel()
|
|
|
|
|
|
|
|
fixturePath := filepath.Join("testdata", "test-provider")
|
|
|
|
tf := e2e.NewBinary(terraformBin, fixturePath)
|
|
|
|
defer tf.Close()
|
|
|
|
|
|
|
|
reattachCh := make(chan *plugin.ReattachConfig)
|
|
|
|
closeCh := make(chan struct{})
|
|
|
|
provider := &providerServer{
|
|
|
|
GRPCProviderServer: grpcplugin.NewGRPCProviderServerShim(test.Provider()),
|
|
|
|
}
|
|
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
|
|
defer cancel()
|
|
|
|
go plugin.Serve(&plugin.ServeConfig{
|
|
|
|
Logger: hclog.New(&hclog.LoggerOptions{
|
|
|
|
Name: "plugintest",
|
|
|
|
Level: hclog.Trace,
|
|
|
|
Output: ioutil.Discard,
|
|
|
|
}),
|
|
|
|
Test: &plugin.ServeTestConfig{
|
|
|
|
Context: ctx,
|
|
|
|
ReattachConfigCh: reattachCh,
|
|
|
|
CloseCh: closeCh,
|
|
|
|
},
|
|
|
|
GRPCServer: plugin.DefaultGRPCServer,
|
|
|
|
VersionedPlugins: map[int]plugin.PluginSet{
|
|
|
|
5: plugin.PluginSet{
|
|
|
|
"provider": &tfplugin.GRPCProviderPlugin{
|
|
|
|
GRPCProvider: func() proto.ProviderServer {
|
|
|
|
return provider
|
|
|
|
},
|
|
|
|
},
|
|
|
|
},
|
|
|
|
},
|
|
|
|
})
|
|
|
|
config := <-reattachCh
|
|
|
|
if config == nil {
|
|
|
|
t.Fatalf("no reattach config received")
|
|
|
|
}
|
|
|
|
reattachStr, err := json.Marshal(map[string]reattachConfig{
|
|
|
|
"hashicorp/test": reattachConfig{
|
|
|
|
Protocol: string(config.Protocol),
|
|
|
|
Pid: config.Pid,
|
|
|
|
Test: true,
|
|
|
|
Addr: reattachConfigAddr{
|
|
|
|
Network: config.Addr.Network(),
|
|
|
|
String: config.Addr.String(),
|
|
|
|
},
|
|
|
|
},
|
|
|
|
})
|
|
|
|
tf.AddEnv("TF_REATTACH_PROVIDERS=" + string(reattachStr))
|
|
|
|
tf.AddEnv("PLUGIN_PROTOCOL_VERSION=5")
|
|
|
|
|
|
|
|
//// INIT
|
|
|
|
stdout, stderr, err := tf.Run("init")
|
|
|
|
if err != nil {
|
|
|
|
t.Fatalf("unexpected init error: %s\nstderr:\n%s", err, stderr)
|
|
|
|
}
|
|
|
|
|
|
|
|
// Make sure we didn't download the binary
|
|
|
|
if strings.Contains(stdout, "Installing hashicorp/test v") {
|
|
|
|
t.Errorf("test provider download message is present in init output:\n%s", stdout)
|
|
|
|
}
|
|
|
|
if tf.FileExists(filepath.Join(".terraform", "plugins", "registry.terraform.io", "hashicorp", "test")) {
|
|
|
|
t.Errorf("test provider binary found in .terraform dir")
|
|
|
|
}
|
|
|
|
|
|
|
|
//// PLAN
|
|
|
|
_, stderr, err = tf.Run("plan", "-out=tfplan")
|
|
|
|
if err != nil {
|
|
|
|
t.Fatalf("unexpected plan error: %s\nstderr:\n%s", err, stderr)
|
|
|
|
}
|
|
|
|
|
2020-09-30 20:28:02 +02:00
|
|
|
if !provider.PlanResourceChangeCalled() {
|
command: Unmanaged providers
This adds supports for "unmanaged" providers, or providers with process
lifecycles not controlled by Terraform. These providers are assumed to
be started before Terraform is launched, and are assumed to shut
themselves down after Terraform has finished running.
To do this, we must update the go-plugin dependency to v1.3.0, which
added support for the "test mode" plugin serving that powers all this.
As a side-effect of not needing to manage the process lifecycle anymore,
Terraform also no longer needs to worry about the provider's binary, as
it won't be used for anything anymore. Because of this, we can disable
the init behavior that concerns itself with downloading that provider's
binary, checking its version, and otherwise managing the binary.
This is all managed on a per-provider basis, so managed providers that
Terraform downloads, starts, and stops can be used in the same commands
as unmanaged providers. The TF_REATTACH_PROVIDERS environment variable
is added, and is a JSON encoding of the provider's address to the
information we need to connect to it.
This change enables two benefits: first, delve and other debuggers can
now be attached to provider server processes, and Terraform can connect.
This allows for attaching debuggers to provider processes, which before
was difficult to impossible. Second, it allows the SDK test framework to
host the provider in the same process as the test driver, while running
a production Terraform binary against the provider. This allows for Go's
built-in race detector and test coverage tooling to work as expected in
provider tests.
Unmanaged providers are expected to work in the exact same way as
managed providers, with one caveat: Terraform kills provider processes
and restarts them once per graph walk, meaning multiple times during
most Terraform CLI commands. As unmanaged providers can't be killed by
Terraform, and have no visibility into graph walks, unmanaged providers
are likely to have differences in how their global mutable state behaves
when compared to managed providers. Namely, unmanaged providers are
likely to retain global state when managed providers would have reset
it. Developers relying on global state should be aware of this.
2020-05-27 02:48:57 +02:00
|
|
|
t.Error("PlanResourceChange not called on in-process provider")
|
|
|
|
}
|
|
|
|
|
|
|
|
//// APPLY
|
|
|
|
_, stderr, err = tf.Run("apply", "tfplan")
|
|
|
|
if err != nil {
|
|
|
|
t.Fatalf("unexpected apply error: %s\nstderr:\n%s", err, stderr)
|
|
|
|
}
|
|
|
|
|
2020-09-30 20:28:02 +02:00
|
|
|
if !provider.ApplyResourceChangeCalled() {
|
command: Unmanaged providers
This adds supports for "unmanaged" providers, or providers with process
lifecycles not controlled by Terraform. These providers are assumed to
be started before Terraform is launched, and are assumed to shut
themselves down after Terraform has finished running.
To do this, we must update the go-plugin dependency to v1.3.0, which
added support for the "test mode" plugin serving that powers all this.
As a side-effect of not needing to manage the process lifecycle anymore,
Terraform also no longer needs to worry about the provider's binary, as
it won't be used for anything anymore. Because of this, we can disable
the init behavior that concerns itself with downloading that provider's
binary, checking its version, and otherwise managing the binary.
This is all managed on a per-provider basis, so managed providers that
Terraform downloads, starts, and stops can be used in the same commands
as unmanaged providers. The TF_REATTACH_PROVIDERS environment variable
is added, and is a JSON encoding of the provider's address to the
information we need to connect to it.
This change enables two benefits: first, delve and other debuggers can
now be attached to provider server processes, and Terraform can connect.
This allows for attaching debuggers to provider processes, which before
was difficult to impossible. Second, it allows the SDK test framework to
host the provider in the same process as the test driver, while running
a production Terraform binary against the provider. This allows for Go's
built-in race detector and test coverage tooling to work as expected in
provider tests.
Unmanaged providers are expected to work in the exact same way as
managed providers, with one caveat: Terraform kills provider processes
and restarts them once per graph walk, meaning multiple times during
most Terraform CLI commands. As unmanaged providers can't be killed by
Terraform, and have no visibility into graph walks, unmanaged providers
are likely to have differences in how their global mutable state behaves
when compared to managed providers. Namely, unmanaged providers are
likely to retain global state when managed providers would have reset
it. Developers relying on global state should be aware of this.
2020-05-27 02:48:57 +02:00
|
|
|
t.Error("ApplyResourceChange not called on in-process provider")
|
|
|
|
}
|
2020-09-30 20:28:02 +02:00
|
|
|
provider.ResetApplyResourceChangeCalled()
|
command: Unmanaged providers
This adds supports for "unmanaged" providers, or providers with process
lifecycles not controlled by Terraform. These providers are assumed to
be started before Terraform is launched, and are assumed to shut
themselves down after Terraform has finished running.
To do this, we must update the go-plugin dependency to v1.3.0, which
added support for the "test mode" plugin serving that powers all this.
As a side-effect of not needing to manage the process lifecycle anymore,
Terraform also no longer needs to worry about the provider's binary, as
it won't be used for anything anymore. Because of this, we can disable
the init behavior that concerns itself with downloading that provider's
binary, checking its version, and otherwise managing the binary.
This is all managed on a per-provider basis, so managed providers that
Terraform downloads, starts, and stops can be used in the same commands
as unmanaged providers. The TF_REATTACH_PROVIDERS environment variable
is added, and is a JSON encoding of the provider's address to the
information we need to connect to it.
This change enables two benefits: first, delve and other debuggers can
now be attached to provider server processes, and Terraform can connect.
This allows for attaching debuggers to provider processes, which before
was difficult to impossible. Second, it allows the SDK test framework to
host the provider in the same process as the test driver, while running
a production Terraform binary against the provider. This allows for Go's
built-in race detector and test coverage tooling to work as expected in
provider tests.
Unmanaged providers are expected to work in the exact same way as
managed providers, with one caveat: Terraform kills provider processes
and restarts them once per graph walk, meaning multiple times during
most Terraform CLI commands. As unmanaged providers can't be killed by
Terraform, and have no visibility into graph walks, unmanaged providers
are likely to have differences in how their global mutable state behaves
when compared to managed providers. Namely, unmanaged providers are
likely to retain global state when managed providers would have reset
it. Developers relying on global state should be aware of this.
2020-05-27 02:48:57 +02:00
|
|
|
|
|
|
|
//// DESTROY
|
|
|
|
_, stderr, err = tf.Run("destroy", "-auto-approve")
|
|
|
|
if err != nil {
|
|
|
|
t.Fatalf("unexpected destroy error: %s\nstderr:\n%s", err, stderr)
|
|
|
|
}
|
|
|
|
|
2020-09-30 20:28:02 +02:00
|
|
|
if !provider.ApplyResourceChangeCalled() {
|
command: Unmanaged providers
This adds supports for "unmanaged" providers, or providers with process
lifecycles not controlled by Terraform. These providers are assumed to
be started before Terraform is launched, and are assumed to shut
themselves down after Terraform has finished running.
To do this, we must update the go-plugin dependency to v1.3.0, which
added support for the "test mode" plugin serving that powers all this.
As a side-effect of not needing to manage the process lifecycle anymore,
Terraform also no longer needs to worry about the provider's binary, as
it won't be used for anything anymore. Because of this, we can disable
the init behavior that concerns itself with downloading that provider's
binary, checking its version, and otherwise managing the binary.
This is all managed on a per-provider basis, so managed providers that
Terraform downloads, starts, and stops can be used in the same commands
as unmanaged providers. The TF_REATTACH_PROVIDERS environment variable
is added, and is a JSON encoding of the provider's address to the
information we need to connect to it.
This change enables two benefits: first, delve and other debuggers can
now be attached to provider server processes, and Terraform can connect.
This allows for attaching debuggers to provider processes, which before
was difficult to impossible. Second, it allows the SDK test framework to
host the provider in the same process as the test driver, while running
a production Terraform binary against the provider. This allows for Go's
built-in race detector and test coverage tooling to work as expected in
provider tests.
Unmanaged providers are expected to work in the exact same way as
managed providers, with one caveat: Terraform kills provider processes
and restarts them once per graph walk, meaning multiple times during
most Terraform CLI commands. As unmanaged providers can't be killed by
Terraform, and have no visibility into graph walks, unmanaged providers
are likely to have differences in how their global mutable state behaves
when compared to managed providers. Namely, unmanaged providers are
likely to retain global state when managed providers would have reset
it. Developers relying on global state should be aware of this.
2020-05-27 02:48:57 +02:00
|
|
|
t.Error("ApplyResourceChange (destroy) not called on in-process provider")
|
|
|
|
}
|
|
|
|
cancel()
|
|
|
|
<-closeCh
|
|
|
|
}
|