terraform/vendor/google.golang.org/grpc/picker_wrapper.go

230 lines
6.3 KiB
Go
Raw Normal View History

/*
*
* Copyright 2017 gRPC authors.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*
*/
package grpc
import (
"context"
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
"fmt"
"io"
"sync"
"google.golang.org/grpc/balancer"
"google.golang.org/grpc/codes"
"google.golang.org/grpc/grpclog"
"google.golang.org/grpc/internal/channelz"
"google.golang.org/grpc/internal/transport"
"google.golang.org/grpc/status"
)
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
// v2PickerWrapper wraps a balancer.Picker while providing the
// balancer.V2Picker API. It requires a pickerWrapper to generate errors
// including the latest connectionError. To be deleted when balancer.Picker is
// updated to the balancer.V2Picker API.
type v2PickerWrapper struct {
picker balancer.Picker
connErr *connErr
}
func (v *v2PickerWrapper) Pick(info balancer.PickInfo) (balancer.PickResult, error) {
sc, done, err := v.picker.Pick(info.Ctx, info)
if err != nil {
if err == balancer.ErrTransientFailure {
return balancer.PickResult{}, balancer.TransientFailureError(fmt.Errorf("%v, latest connection error: %v", err, v.connErr.connectionError()))
}
return balancer.PickResult{}, err
}
return balancer.PickResult{SubConn: sc, Done: done}, nil
}
// pickerWrapper is a wrapper of balancer.Picker. It blocks on certain pick
// actions and unblock when there's a picker update.
type pickerWrapper struct {
mu sync.Mutex
done bool
blockingCh chan struct{}
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
picker balancer.V2Picker
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
// The latest connection error. TODO: remove when V1 picker is deprecated;
// balancer should be responsible for providing the error.
*connErr
}
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
type connErr struct {
mu sync.Mutex
err error
}
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 (c *connErr) updateConnectionError(err error) {
c.mu.Lock()
c.err = err
c.mu.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
func (c *connErr) connectionError() error {
c.mu.Lock()
err := c.err
c.mu.Unlock()
return err
}
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 newPickerWrapper() *pickerWrapper {
return &pickerWrapper{blockingCh: make(chan struct{}), connErr: &connErr{}}
}
// updatePicker is called by UpdateBalancerState. It unblocks all blocked pick.
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 (pw *pickerWrapper) updatePicker(p balancer.Picker) {
pw.updatePickerV2(&v2PickerWrapper{picker: p, connErr: pw.connErr})
}
// updatePicker is called by UpdateBalancerState. It unblocks all blocked pick.
func (pw *pickerWrapper) updatePickerV2(p balancer.V2Picker) {
pw.mu.Lock()
if pw.done {
pw.mu.Unlock()
return
}
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
pw.picker = p
// pw.blockingCh should never be nil.
close(pw.blockingCh)
pw.blockingCh = make(chan struct{})
pw.mu.Unlock()
}
func doneChannelzWrapper(acw *acBalancerWrapper, done func(balancer.DoneInfo)) func(balancer.DoneInfo) {
acw.mu.Lock()
ac := acw.ac
acw.mu.Unlock()
ac.incrCallsStarted()
return func(b balancer.DoneInfo) {
if b.Err != nil && b.Err != io.EOF {
ac.incrCallsFailed()
} else {
ac.incrCallsSucceeded()
}
if done != nil {
done(b)
}
}
}
// pick returns the transport that will be used for the RPC.
// It may block in the following cases:
// - there's no picker
// - the current picker returns ErrNoSubConnAvailable
// - the current picker returns other errors and failfast is false.
// - the subConn returned by the current picker is not READY
// When one of these situations happens, pick blocks until the picker gets updated.
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 (pw *pickerWrapper) pick(ctx context.Context, failfast bool, info balancer.PickInfo) (transport.ClientTransport, func(balancer.DoneInfo), error) {
var ch chan struct{}
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
var lastPickErr error
for {
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
pw.mu.Lock()
if pw.done {
pw.mu.Unlock()
return nil, nil, ErrClientConnClosing
}
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
if pw.picker == nil {
ch = pw.blockingCh
}
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
if ch == pw.blockingCh {
// This could happen when either:
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
// - pw.picker is nil (the previous if condition), or
// - has called pick on the current picker.
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
pw.mu.Unlock()
select {
case <-ctx.Done():
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
var errStr string
if lastPickErr != nil {
errStr = "latest balancer error: " + lastPickErr.Error()
} else if connectionErr := pw.connectionError(); connectionErr != nil {
errStr = "latest connection error: " + connectionErr.Error()
} else {
errStr = ctx.Err().Error()
}
switch ctx.Err() {
case context.DeadlineExceeded:
return nil, nil, status.Error(codes.DeadlineExceeded, errStr)
case context.Canceled:
return nil, nil, status.Error(codes.Canceled, errStr)
2019-09-09 14:04:58 +02:00
}
case <-ch:
}
continue
}
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
ch = pw.blockingCh
p := pw.picker
pw.mu.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
pickResult, err := p.Pick(info)
if err != nil {
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
if err == balancer.ErrNoSubConnAvailable {
continue
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
}
if tfe, ok := err.(interface{ IsTransientFailure() bool }); ok && tfe.IsTransientFailure() {
if !failfast {
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
lastPickErr = err
continue
}
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
return nil, nil, status.Error(codes.Unavailable, err.Error())
}
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
if _, ok := status.FromError(err); ok {
return nil, nil, err
}
// err is some other error.
return nil, nil, status.Error(codes.Unknown, err.Error())
}
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
acw, ok := pickResult.SubConn.(*acBalancerWrapper)
if !ok {
grpclog.Error("subconn returned from pick is not *acBalancerWrapper")
continue
}
if t, ok := acw.getAddrConn().getReadyTransport(); ok {
if channelz.IsOn() {
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
return t, doneChannelzWrapper(acw, pickResult.Done), nil
}
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
return t, pickResult.Done, nil
}
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
if pickResult.Done != nil {
2019-09-09 14:04:58 +02:00
// Calling done with nil error, no bytes sent and no bytes received.
// DoneInfo with default value works.
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
pickResult.Done(balancer.DoneInfo{})
2019-09-09 14:04:58 +02:00
}
grpclog.Infof("blockingPicker: the picked transport is not ready, loop back to repick")
// If ok == false, ac.state is not READY.
// A valid picker always returns READY subConn. This means the state of ac
// just changed, and picker will be updated shortly.
// continue back to the beginning of the for loop to repick.
}
}
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 (pw *pickerWrapper) close() {
pw.mu.Lock()
defer pw.mu.Unlock()
if pw.done {
return
}
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
pw.done = true
close(pw.blockingCh)
}