* Add a list of in-flight operations to the deployment
This commit augments 'DeploymentV2' with a list of operations that are
currently in flight. This information is used by the engine to keep
track of whether or not a particular deployment is in a valid state.
The SnapshotManager is responsible for inserting and removing operations
from the in-flight operation list. When the engine registers an intent
to perform an operation, SnapshotManager inserts an Operation into this
list and saves it to the snapshot. When an operation completes, the
SnapshotManager removes it from the snapshot. From this, the engine can
infer that if it ever sees a deployment with pending operations, the
Pulumi CLI must have crashed or otherwise abnormally terminated before
seeing whether or not an operation completed successfully.
To remedy this state, this commit also adds code to 'pulumi stack
import' that clears all pending operations from a deployment, as well as
code to plan generation that will reject any deployments that have
pending operations present.
At the CLI level, if we see that we are in a state where pending
operations were in-flight when the engine died, we'll issue a
human-friendly error message that indicates which resources are in a bad
state and how to recover their stack.
* CR: Multi-line string literals, renaming in-flight -> pending
* CR: Add enum to apitype for operation type, also name status -> type for clarity
* Fix the yaml type
* Fix missed renames
* Add implementation for lifecycle_test.go
* Rebase against master
* Initial support for passing URLs to `new` and `up`
This PR adds initial support for `pulumi new` using Git under the covers
to manage Pulumi templates, providing the same experience as before.
You can now also optionally pass a URL to a Git repository, e.g.
`pulumi new [<url>]`, including subdirectories within the repository,
and arbitrary branches, tags, or commits.
The following commands result in the same behavior from the user's
perspective:
- `pulumi new javascript`
- `pulumi new https://github.com/pulumi/templates/templates/javascript`
- `pulumi new https://github.com/pulumi/templates/tree/master/templates/javascript`
- `pulumi new https://github.com/pulumi/templates/tree/HEAD/templates/javascript`
To specify an arbitrary branch, tag, or commit:
- `pulumi new https://github.com/pulumi/templates/tree/<branch>/templates/javascript`
- `pulumi new https://github.com/pulumi/templates/tree/<tag>/templates/javascript`
- `pulumi new https://github.com/pulumi/templates/tree/<commit>/templates/javascript`
Branches and tags can include '/' separators, and `pulumi` will still
find the right subdirectory.
URLs to Gists are also supported, e.g.:
`pulumi new https://gist.github.com/justinvp/6673959ceb9d2ac5a14c6d536cb871a6`
If the specified subdirectory in the repository does not contain a
`Pulumi.yaml`, it will look for subdirectories within containing
`Pulumi.yaml` files, and prompt the user to choose a template, along the
lines of how `pulumi new` behaves when no template is specified.
The following commands result in the CLI prompting to choose a template:
- `pulumi new`
- `pulumi new https://github.com/pulumi/templates/templates`
- `pulumi new https://github.com/pulumi/templates/tree/master/templates`
- `pulumi new https://github.com/pulumi/templates/tree/HEAD/templates`
Of course, arbitrary branches, tags, or commits can be specified as well:
- `pulumi new https://github.com/pulumi/templates/tree/<branch>/templates`
- `pulumi new https://github.com/pulumi/templates/tree/<tag>/templates`
- `pulumi new https://github.com/pulumi/templates/tree/<commit>/templates`
This PR also includes initial support for passing URLs to `pulumi up`,
providing a streamlined way to deploy installable cloud applications
with Pulumi, without having to manage source code locally before doing
a deployment.
For example, `pulumi up https://github.com/justinvp/aws` can be used to
deploy a sample AWS app. The stack can be updated with different
versions, e.g.
`pulumi up https://github.com/justinvp/aws/tree/v2 -s <stack-to-update>`
Config values can optionally be passed via command line flags, e.g.
`pulumi up https://github.com/justinvp/aws -c aws:region=us-west-2 -c foo:bar=blah`
Gists can also be used, e.g.
`pulumi up https://gist.github.com/justinvp/62fde0463f243fcb49f5a7222e51bc76`
* Fix panic when hitting ^C from "choose template" prompt
* Add description to templates
When running `pulumi new` without specifying a template, include the template description along with the name in the "choose template" display.
```
$ pulumi new
Please choose a template:
aws-go A minimal AWS Go program
aws-javascript A minimal AWS JavaScript program
aws-python A minimal AWS Python program
aws-typescript A minimal AWS TypeScript program
> go A minimal Go program
hello-aws-javascript A simple AWS serverless JavaScript program
javascript A minimal JavaScript program
python A minimal Python program
typescript A minimal TypeScript program
```
* React to changes to the pulumi/templates repo.
We restructured the `pulumi/templates` repo to have all the templates in the root instead of in a `templates` subdirectory, so make the change here to no longer look for templates in `templates`.
This also fixes an issue around using `Depth: 1` that I found while testing this. When a named template is used, we attempt to clone or pull from the `pulumi/templates` repo to `~/.pulumi/templates`. Having it go in this well-known directory allows us to maintain previous behavior around allowing offline use of templates. If we use `Depth: 1` for the initial clone, it will fail when attempting to pull when there are updates to the remote repository. Unfortunately, there's no built-in `--unshallow` support in `go-git` and setting a larger `Depth` doesn't appear to help. There may be a workaround, but for now, if we're cloning the pulumi templates directory to `~/.pulumi/templates`, we won't use `Depth: 1`. For template URLs, we will continue to use `Depth: 1` as we clone those to a temp directory (which gets deleted) that we'll never try to update.
* List available templates in help text
* Address PR Feedback
* Don't show "Installing dependencies" message for `up`
* Fix secrets handling
When prompting for config, if the existing stack value is a secret, keep it a secret and mask the prompt. If the template says it should be secret, make it a secret.
* Fix ${PROJECT} and ${DESCRIPTION} handling for `up`
Templates used with `up` should already have a filled-in project name and description, but if it's a `new`-style template, that has `${PROJECT}` and/or `${DESCRIPTION}`, be helpful and just replace these with better values.
* Fix stack handling
Add a bool `setCurrent` param to `requireStack` to control whether the current stack should be saved in workspace settings. For the `up <url>` case, we don't want to save. Also, split the `up` code into two separate functions: one for the `up <url>` case and another for the normal `up` case where you have workspace in your current directory. While we may be able to combine them back into a single function, right now it's a bit cleaner being separate, even with some small amount of duplication.
* Fix panic due to nil crypter
Lazily get the crypter only if needed inside `promptForConfig`.
* Embellish comment
* Harden isPreconfiguredEmptyStack check
Fix the code to check to make sure the URL specified on the command line matches the URL stored in the `pulumi:template` config value, and that the rest of the config from the stack satisfies the config requirements of the template.
Many non-secrets are actually pretty high entropy, at least according
to `zxcvbn`. For example: "Hello, Pulumi Timers!" would actually cause
us to say: "this looks like a secret", much in the same way that
"correct horse battery staple" is high entropy according to that
package.
In addition to considering the entropy of the value, gosec (the linter
we copied this logic from) also considers the name of the value that
is being assigned to.
In that spirit, let's only do this check when the config key name
actually looks like it is something we'd want to warn the user about.
We use the same regular expression as `gosec`.
Fixes#1732
In #1341 we promoted a class of errors in fetching git metadata from
glog messages to warnings printed by the CLI. On the asumption that
when we got warnings here they would be actionable.
The major impact here is that when you are working in a repository
which does not have a remote set to GitHub (common if you have just
`git init`'d a repository for a new project) or you don't call your
remote `origin` or you use some other code provider, we end up
printing a warning during every update.
This change does two things:
- Restructure the way we detect metadata to attempt to make progress
when it can. We bias towards returning some metadata even when we
can't determine the complete set of metadata.
- Use a multierror to track all the underlying failures from our
metadata probing and move it back to a glog message.
Overall, this feels like the right balance to me. We are retaining the
rich diagnostics information for when things go wrong, but we aren't
warning about common cases.
We could, of course, try to tighten our huristics (e.g. don't warn if
we can't find a GitHub remote but do warn if we can't compute if the
worktree is dirty) but it feels like that will be a game of
whack-a-mole over time and when warnings do fire its unlikely they
will be actionable.
Fixes#1443
### First-Class Providers
These changes implement support for first-class providers. First-class
providers are provider plugins that are exposed as resources via the
Pulumi programming model so that they may be explicitly and multiply
instantiated. Each instance of a provider resource may be configured
differently, and configuration parameters may be source from the
outputs of other resources.
### Provider Plugin Changes
In order to accommodate the need to verify and diff provider
configuration and configure providers without complete configuration
information, these changes adjust the high-level provider plugin
interface. Two new methods for validating a provider's configuration
and diffing changes to the same have been added (`CheckConfig` and
`DiffConfig`, respectively), and the type of the configuration bag
accepted by `Configure` has been changed to a `PropertyMap`.
These changes have not yet been reflected in the provider plugin gRPC
interface. We will do this in a set of follow-up changes. Until then,
these methods are implemented by adapters:
- `CheckConfig` validates that all configuration parameters are string
or unknown properties. This is necessary because existing plugins
only accept string-typed configuration values.
- `DiffConfig` either returns "never replace" if all configuration
values are known or "must replace" if any configuration value is
unknown. The justification for this behavior is given
[here](https://github.com/pulumi/pulumi/pull/1695/files#diff-a6cd5c7f337665f5bb22e92ca5f07537R106)
- `Configure` converts the config bag to a legacy config map and
configures the provider plugin if all config values are known. If any
config value is unknown, the underlying plugin is not configured and
the provider may only perform `Check`, `Read`, and `Invoke`, all of
which return empty results. We justify this behavior becuase it is
only possible during a preview and provides the best experience we
can manage with the existing gRPC interface.
### Resource Model Changes
Providers are now exposed as resources that participate in a stack's
dependency graph. Like other resources, they are explicitly created,
may have multiple instances, and may have dependencies on other
resources. Providers are referred to using provider references, which
are a combination of the provider's URN and its ID. This design
addresses the need during a preview to refer to providers that have not
yet been physically created and therefore have no ID.
All custom resources that are not themselves providers must specify a
single provider via a provider reference. The named provider will be
used to manage that resource's CRUD operations. If a resource's
provider reference changes, the resource must be replaced. Though its
URN is not present in the resource's dependency list, the provider
should be treated as a dependency of the resource when topologically
sorting the dependency graph.
Finally, `Invoke` operations must now specify a provider to use for the
invocation via a provider reference.
### Engine Changes
First-class providers support requires a few changes to the engine:
- The engine must have some way to map from provider references to
provider plugins. It must be possible to add providers from a stack's
checkpoint to this map and to register new/updated providers during
the execution of a plan in response to CRUD operations on provider
resources.
- In order to support updating existing stacks using existing Pulumi
programs that may not explicitly instantiate providers, the engine
must be able to manage the "default" providers for each package
referenced by a checkpoint or Pulumi program. The configuration for
a "default" provider is taken from the stack's configuration data.
The former need is addressed by adding a provider registry type that is
responsible for managing all of the plugins required by a plan. In
addition to loading plugins froma checkpoint and providing the ability
to map from a provider reference to a provider plugin, this type serves
as the provider plugin for providers themselves (i.e. it is the
"provider provider").
The latter need is solved via two relatively self-contained changes to
plan setup and the eval source.
During plan setup, the old checkpoint is scanned for custom resources
that do not have a provider reference in order to compute the set of
packages that require a default provider. Once this set has been
computed, the required default provider definitions are conjured and
prepended to the checkpoint's resource list. Each resource that
requires a default provider is then updated to refer to the default
provider for its package.
While an eval source is running, each custom resource registration,
resource read, and invoke that does not name a provider is trapped
before being returned by the source iterator. If no default provider
for the appropriate package has been registered, the eval source
synthesizes an appropriate registration, waits for it to complete, and
records the registered provider's reference. This reference is injected
into the original request, which is then processed as usual. If a
default provider was already registered, the recorded reference is
used and no new registration occurs.
### SDK Changes
These changes only expose first-class providers from the Node.JS SDK.
- A new abstract class, `ProviderResource`, can be subclassed and used
to instantiate first-class providers.
- A new field in `ResourceOptions`, `provider`, can be used to supply
a particular provider instance to manage a `CustomResource`'s CRUD
operations.
- A new type, `InvokeOptions`, can be used to specify options that
control the behavior of a call to `pulumi.runtime.invoke`. This type
includes a `provider` field that is analogous to
`ResourceOptions.provider`.
This change lets us set runtime specific options in Pulumi.yaml, which
will flow as arguments to the language hosts. We then teach the nodejs
host that when the `typescript` is set to `true` that it should load
ts-node before calling into user code. This allows using typescript
natively without an explicit compile step outside of Pulumi.
This works even when a tsconfig.json file is not present in the
application and should provide a nicer inner loop for folks writing
typescript (I'm pretty sure everyone has run into the "but I fixed
that bug! Why isn't it getting picked up? Oh, I forgot to run tsc"
problem.
Fixes#958
* Protobuf changes to record dependencies for read resources
* Add a number of tests for read resources, especially around replacement
* Place read resources in the snapshot with "external" bit set
Fixespulumi/pulumi#1521. This commit introduces two new step ops: Read
and ReadReplacement. The engine generates Read and ReadReplacement steps
when servicing ReadResource RPC calls from the language host.
* Fix an omission of OpReadReplace from the step list
* Rebase against master
* Transition to use V2 Resources by default
* Add a semantic "relinquish" operation to the engine
If the engine observes that a resource is read and also that the
resource exists in the snapshot as a non-external resource, it will not
delete the resource if the IDs of the old and new resources match.
* Typo fix
* CR: add missing comments, DeserializeDeployment -> DeserializeDeploymentV2, ID check
Previously, we would unconditionally warn anytime you added a non-secret
config:
$ pulumi config set aws:region us-west-2
warning: saved config key '%s' value '%s' as plaintext;
re-run with --secret to encrypt the value instead.
Use --plaintext to avoid this warning
This was particularly annoying, since it is very common to store
non-secret config. For instance, the AWS region. And it was easy to tune
out because it wasn't actually warning about anything interesting.
This change, which resolvespulumi/pulumi#570, uses an approach similar
to Go's gas linter, to detect high entropy values, and issue an error.
This ensures that we only make noise on things we suspect are actually
secrets being stored in plaintext, and forces the user to pass
--plaintext. For instance, the common case issues no errors:
$ pulumi config set aws:region us-west-2
And in the event that you store something that is secret-like:
$ pulumi config set aws:region nq8r4B4xslzrtj0a3
error: config value 'nq8r4B4xslzrtj0a3' looks like a secret;
rerun with --secret to encrypt it, or --plaintext if you meant
to store in plaintext
To suppress this, simply pass --secret (to encrypt) or --plaintext (to
override the warning).
After running `pulumi new`, we print a message to let the user
know the project has been created, along with next steps to actually
perform a deployment. It's easy for this to get lost among the rest
of the output, however. So, wordsmith it a little, and add some color,
to make it pop a little bit more.
Also add SuggestFor annotations so that `init` and `create` direct
you to run the `new` command (I often accidentally type `init`).
Previously, it would only show the first 7 templates (we currently have 9 total), and you'd have to move the cursor down to the bottom to show the last 2.
Instead of needing you to first select the current stack to use any of
these commands, allow passing `-s <stack-name>` or `--stack
<stack-name>` to say what stack you want to operate on.
These commands still require a `Pulumi.yaml` file to be present, which
is not ideal, but would require a larger refactoring to fix. That
refactoring will happen as part of #1556.
Fixes#1370
This was an artifact of history. Since we'll be supporting the local
backend, we don't need yet another flag guarding it (you already have
to opt in with -c local:// which is enough of a hoop).
This change fixes the following warning issued by using `go get` with
the Pulumi repository:
package github.com/moby/moby/api/types/time: code in directory
/Users/James/Code/go/src/github.com/moby/moby/api/types/time expects
import "github.com/docker/docker/api/types/time"
Moby defines canonical import paths [1] reflecting the lineage back to
the `docker/docker` repository.
[1]: https://github.com/moby/moby/blob/master/api/types/time/duration_convert.go#L1
Usage:
```
pulumi new <template> --dir folderName
```
Used to specify the directory where to place the generated project.
If the directory does not exist, it will be created.
* Error when loading a deployment that is not a version that the CLI understands
* Add a test for 'pulumi stack import' on a badly-versioned deployment
* Move current deployment version to 'apitype'
* Rebase against master
* CR: emit CLI-friendly error message at the two points outside of the engine calling 'DeserializeDeployment'
This change supports displaying manifest information for a stack and
changes the way we handle Snapshots in our backend.
Previously, every call to GetStack would synthesize a Snapshot by
taking the set of resources returned from the
`/api/stacks/<owner>/<name>` endpoint, combined with an empty
manfiest (since the service was not returning the manifest).
This wasn't great for two reasons:
1. We didn't have manifest information, so we couldn't display any of
its information (most important the last updated time).
2. This strategy required that the service return all the resources
for a stack anytime GetStack was called. While the CLI did not
often need this detailed information the fact that we forced the
Service to produce it (which in the case of stack managed PPC would
require the service to talk to yet another service) creates a bunch
of work that we end up ignoring.
I've refactored the code such that `backend.Stack`'s `Snapshot()` method
now lazily requests the information from the service such that we can
construct a `Snapshot()` on demand and only pay the cost when we
actually need it.
I think making more of this stuff lazy is the long term direction we
want to follow.
Unfortunately, right now, it means in cases where we do need this data
we end up fetching it twice. The service does it once when we call
GetStack and then we do it again when we actually need to get at the
Snapshot. However, once we land this change, we can update the
service to no longer return resources on the apistack.Stack type. The
CLI no longer needs this property. We'll likely want to continue in a
direction where `apistack.Stack` can be created quickly by the
service (without expensive database queries or fetching remote
resources) and just add additional endpoints that let us get at the
specific information we want in the specific cases when we want it
instead of forcing us to return a bunch of data that we often ignore.
Fixespulumi/pulumi-service#371
While we wait for upstream to fix the security issue, we'll just pass
`--no-audit` to `npm install` when creating a new project. Since the
warning is against an indirect dependency of @pulumi/pulumi, we can't
actually address the issue ourselves.
Mitigates #1350
In pulumi/pulumi#1356, we observed that we can fail during a destroy
because we attempt to load the language plugin, which now eagerly looks
for the @pulumi/pulumi package.
This is also blocking ingestion of the latest engine bits into the PPC.
It turns out that for destroy (and refresh), we have no need for the
language plugin. So, let's skip loading it when appropriate.
These changes add support for adding a tracing header to API requests
made to the Pulumi service. Setting the `PULUMI_TRACING_HEADER`
environment variable or enabling debug commands and passing the
`--tracing-header` will change the value sent in this header. Setting
this value to `1` will request that the service enable distributed
tracing for all requests made by a particular CLI invocation.
The newly added `pulumi config refresh` updates your local copy of the
Pulumi.<stack-name>.yaml file to have the same configuration as the
most recent deployment in the cloud.
This can be used in a varirty of ways. One place we plan to use it is
in automation to clean up "leaked" stacks we have in CI. With the
changes you'll now be able to do the following:
```
$ cd $(mktemp -d)
$ echo -e "name: who-cares\nruntime: nodejs" > Pulumi.yaml
$ pulumi stack select <leaked-stack-name>
$ pulumi config refresh -f
$ pulumi destroy --force
```
Having a simpler gesture for the above is something we'll want to do
long term (we should be able to support `pulumi destory <stack-name>`
from a completely empty folder, today you need a Pulumi.yaml file
present, even if the contents don't matter).
But this gets us a little closer to where we want to be and introduces
a helpful primitive in the system.
Contributes to #814
These changes add support for injecting client tracing spans into HTTP
requests to the Pulumi API. The server can then rematerialize these span
references and attach its own spans for distributed tracing.
This change moves Git failures from glog.Warnings, which we don't
really pay attention, to true CLI warnings.
This will ensure we at least get bug reports in the event that this
fails on some user machine out in the wild. They are still non-fatal,
of course, since such a failure needn't prevent an update from happening.
This change captures the Git committer and author's login and email
addresses, so that we can display them prominently in the service. At
the moment, we only attribute updates to the identity that performed the
update which, in CI scenarios, is often always the same person for an
organization. This makes Pulumi look like a needlessly lonely place.
These changes enable tracing of Pulumi API calls.
The span with which to associate an API call is passed via a
`context.Context` parameter. This required plumbing a
`context.Context` parameter through a rather large number of APIs,
especially in the backend.
In general, all API calls are associated with a new root span that
exists for essentially the entire lifetime of an invocation of the
Pulumi CLI. There were a few places where the plumbing got a bit hairier
than I was willing to address with these changes; I've used
`context.Background()` in these instances. API calls that receive this
context will create new root spans, but will still be traced.
* Initialize a new stack as part of `pulumi new`
* Prompt for values with defaults preselected
* Install dependencies
* Prompt for default config values
This changes the CLI interface in a few ways:
* `pulumi preview` is back! The alternative of saying
`pulumi update --preview` just felt awkward, and it's a common
operation to want to perform. Let's just make it work.
* There are two flags consistent across all update commands,
`update`, `refresh`, and `destroy`:
- `--skip-preview` will skip the preview step. Note that this
does *not* skip the prompt to confirm that you'd like to proceed.
Indeed, it will still prompt, with a little warning text about
the fact that the preview has been skipped.
* `--yes` will auto-approve the updates.
This lands us in a simpler and more intuitive spot for common scenarios.