* [Parallelism] Introduce a "step generator" component by refactoring all
step generation logic out of PlanIterator
* CR: remove dead fields on PlanIterator
* Delete Before Create
This commit implements the full semantics of delete before create. If a
resource is replaced and requires deletion before creation, the engine
will use the dependency graph saved in the snapshot to delete all
resources that depend on the resource being replaced prior to the
deletion of the resource to be replaced.
* Rebase against master
* CR: Simplify the control flow in makeRegisterResourceSteps
* Run Check on new inputs when re-creating a resource
* Fix an issue where the planner emitted benign but incorrect deletes of DBR-deleted resources
* CR: produce the list of dependent resources in dependency order and iterate over the list in reverse
* CR: deps->dependents, fix an issue with DependingOn where duplicate nodes could be added to the dependent set
* CR: Fix an issue where we were considering old defaults and new inputs
inappropriately when re-creating a deleted resource
* CR: save 'iter.deletes[urn]' as a local, iterate starting at cursorIndex + 1 for dependency graph
* Refactor the SnapshotManager interface
Lift snapshot management out of the engine by delegating it to the
SnapshotManager implementation in pkg/backend.
* Add a event interface for plugin loads and use that interface to record plugins in the snapshot
* Remove dead code
* Add comments to Events
* Add a number of tests for SnapshotManager
* CR feedback: use a successful bit on 'End' instead of having a separate 'Abort' API
* CR feedback
* CR feedback: register plugins one-at-a-time instead of the entire state at once
This issue arises becuase the behavior we're currently getting from Diff
for TF-based providers differs from the behavior we expect. We are
presenting the provider with the old state and new inputs. If the old
state contains output properties that differ from the new inputs, the
provider will detect a diff where we may expect no changes.
Rather than deferring to the provider for all diffs, these changes only
defer to the provider if a legacy diff was detected (i.e. there is some
difference between the old and new provider-calculated inputs).
This change implements a `pulumi refresh` command. It operates a bit
like `pulumi update`, and friends, in that it supports `--preview` and
`--diff`, along with the usual flags, and will update your checkpoint.
It works through substitution of the deploy.Source abstraction, which
generates a sequence of resource registration events. This new
deploy.RefreshSource takes in a prior checkpoint and will walk it,
refreshing the state via the associated resource providers by invoking
Read for each resource encountered, and merging the resulting state with
the prior checkpoint, to yield a new resource.Goal state. This state is
then fed through the engine in the usual ways with a few minor caveats:
namely, although the engine must generate steps for the logical
operations (permitting us to get nice summaries, progress, and diffs),
it mustn't actually carry them out because the state being imported
already reflects reality (a deleted resource has *already* been deleted,
so of course the engine need not perform the deletion). The diffing
logic also needs to know how to treat the case of refresh slightly
differently, because we are going to be diffing outputs and not inputs.
Note that support for managed stacks is not yet complete, since that
requires updates to the service to support a refresh endpoint. That
will be coming soon ...
* Lift snapshot management out of the engine
This PR is a prerequisite for parallelism by addressing a major problem
that the engine has to deal with when performing parallel resource
construction: parallel mutation of the global snapshot. This PR adds
a `SnapshotManager` type that is responsible for maintaining and
persisting the current resource snapshot. It serializes all reads and
writes to the global snapshot and persists the snapshot to persistent
storage upon every write.
As a side-effect of this, the core engine no longer needs to know about
snapshot management at all; all snapshot operations can be handled as
callbacks on deployment events. This will greatly simplify the
parallelization of the core engine.
Worth noting is that the core engine will still need to be able to read
the current snapshot, since it is interested in the dependency graphs
contained within. The full implications of that are out of scope of this
PR.
Remove dead code, Steps no longer need a reference to the plan iterator that created them
Fixing various issues that arise when bringing up pulumi-aws
Line length broke the build
Code review: remove dead field, fix yaml name error
Rebase against master, provide implementation of StackPersister for cloud backend
Code review feedback: comments on MutationStatus, style in snapshot.go
Code review feedback: move SnapshotManager to pkg/backend, change engine to use an interface SnapshotManager
Code review feedback: use a channel for synchronization
Add a comment and a new test
* Maintain two checkpoints, an immutable base and a mutable delta, and
periodically merge the two to produce snapshots
* Add a lot of tests - covers all of the non-error paths of BeginMutation and End
* Fix a test resource provider
* Add a few tests, fix a few issues
* Rebase against master, fixed merge
This commit changes two things about our resource model:
* Stop performing Pulumi Engine-side diffing of resource state.
Instead, we defer to the resource plugins themselves to determine
whether a change was made and, if so, the extent of it. This
manifests as a simple change to the Diff function; it is done in
a backwards compatible way so that we continue with legacy diffing
for existing resource provider plugins.
* Add a Read RPC method for resource providers. It simply takes a
resource's ID and URN, plus an optional bag of further qualifying
state, and it returns the current property state as read back from
the actual live environment. Note that the optional bag of state
must at least include enough additional properties for resources
wherein the ID is insufficient for the provider to perform a lookup.
It may, however, include the full bag of prior state, for instance
in the case of a refresh operation.
This is part of pulumi/pulumi#1108.
* Improve error messages output by the CLI
This fixes a couple known issues with the way that we present errors
from the Pulumi CLI:
1. Any errors from RPC endpoints were bubbling up as they were to
the top-level, which was unfortunate because they contained
RPC-specific noise that we don't want to present to the user. This
commit unwraps errors from resource providers.
2. The "catastrophic error" message often got printed twice
3. Fatal errors are often printed twice, because our CLI top-level
prints out the fatal error that it receives before exiting. A lot of
the time this error has already been printed.
4. Errors were prefixed by PU####.
* Feedback: Omit the 'catastrophic' error message and use a less verbose error message as the final error
* Code review feedback: interpretRPCError -> resourceStateAndError
* Code review feedback: deleting some commented-out code, error capitalization
* Cleanup after rebase
As it stands, we only configure those providers for which configuration
is present. This can lead to surprising failure modes if those providers
are then used to create resources. These changes ensure that all
resource providers that are not configured during plan initialization
are configured upon first load.
Fixes#758.
Most of the errors in this package are holdovers from our previous
syetem where we had our own custom compiler and evaluator and are no
longer needed. The few we still use during plan applicaton (via the
diagnostics system, which is another component from the old system
that we still use) have been promoted into the diag package. Doing so,
allows us to not have to import "github.com/pkg/errors" as "goerr" in
some parts of the engine, a nice cleaup.
This commit does two things:
1. All dependencies of a resource, both implicit and explicit, are
communicated directly to the engine when registering a resource. The
engine keeps track of these dependencies and ultimately serializes
them out to the checkpoint file upon successful deployment.
2. Once a successful deployment is done, the new `pulumi stack
graph` command reads the checkpoint file and outputs the dependency
information within in the DOT format.
Keeping track of dependency information within the checkpoint file is
desirable for a number of reasons, most notably delete-before-create,
where we want to delete resources before we have created their
replacement when performing an update.
This change adds a GetRequiredPlugins RPC method to the language
host, enabling us to query it for its list of plugin requirements.
This is language-specific because it requires looking at the set
of dependencies (e.g., package.json files).
It also adds a call up front during any update/preview operation
to compute the set of plugins and require that they are present.
These plugins are populated in the cache and will be used for all
subsequent plugin-related operations during the engine's activity.
We now cache the language plugins, so that we may load them
eagerly too, which we never did previously due to the fact that
we needed to pass the monitor address at load time. This was a
bit bizarre anyhow, since it's really the Run RPC function that
needs this information. So, to enable caching and eager loading
-- which we need in order to invoke GetRequiredPlugins -- the
"phone home" monitor RPC address is passed at Run time.
In a subsequent change, we will switch to faulting in the plugins
that are missing -- rather than erroring -- in addition to
supporting the `pulumi plugin install` CLI command.
This addresses pulumi/pulumi#446: what we used to call "package" is
now called "project". This has gotten more confusing over time, now
that we're doing real package management.
Also fixespulumi/pulumi#426, while in here.
My previous change to stop supplying unknown properties to providers
broke `pulumi preview` in the case of unknown inputs. This change
restores the previous behavior for previews only; the new unknown-free
behavior remains for applies.
Fixes#790.
This change implements resource protection, as per pulumi/pulumi#689.
The overall idea is that a resource can be marked as "protect: true",
which will prevent deletion of that resource for any reason whatsoever
(straight deletion, replacement, etc). This is expressed in the
program. To "unprotect" a resource, one must perform an update setting
"protect: false", and then afterwards, they can delete the resource.
For example:
let res = new MyResource("precious", { .. }, { protect: true });
Afterwards, the resource will display in the CLI with a lock icon, and
any attempts to remove it will fail in the usual ways (in planning or,
worst case, during an actual update).
This was done by adding a new ResourceOptions bag parameter to the
base Resource types. This is unfortunately a breaking change, but now
is the right time to take this one. We had been adding new settings
one by one -- like parent and dependsOn -- and this new approach will
set us up to add any number of additional settings down the road,
without needing to worry about breaking anything ever again.
This is related to protected stacks, as described in
pulumi/pulumi-service#399. Most likely this will serve as a foundational
building block that enables the coarser grained policy management.
We do not need all of the information in the old state for this call, as
outputs will not be read by the provider during validation or defaults
computation.
This change passes a resource's old output state, so that it contains
everything -- defaults included -- for purposes of the provider's diffing.
Not doing so can lead the provider into thinking some of the requisite
state is missing.
This change adds rudimentary delete-before-create support (see
pulumi/pulumi#450). This cannot possibly be complete until we also
implement pulumi/pulumi#624, becuase we may try to delete a resource
while it still has dependent resources (which almost certainly will
fail). But until then, we can use this to manually unwedge ourselves
for leaf-node resources that do not support old and new resources
living side-by-side.
The two-phase output properties change broke the ability to recover
from a failed replacement that yields pending deletes in the checkpoint.
The issue here is simply that we should remember pending registrations
only for logical operations that *also* have a "new" state (create or
update). This commit fixes this, and also adds a new step test with
fault injection to probe many interesting combinations of steps.
Every single resource has a type prefix of
pulumi:pulumi:Stack$
which makes URNs quite lengthy without adding any value. Since
they all have this prefix, adding it doesn't help to disambiguate.
This change skips adding the parent URN part when it is the built-in
automatic stack type name.
At some point, we fixed a bug in the way state is managed for "same"
steps, which meant that we wouldn't see newly added output properties.
This had the effect that, if you had a stack already stood up, and
updated it to have output properties, we would miss them. (Stacks
stood up from scratch would still have them.) This fixes that problem,
in addition to two other things: 1) we need to sort output property
names to ensure a deterministic ordering, and 2) we need to also
unconditionally apply the outputs RPC coming in, to ensure that the
resulting resource always has the correct outputs (so that for example
deleting prior output properties actually deletes them).
Also add some testing for this area to make sure we don't break again.
Fixespulumi/pulumi#631.
These changes push the `config.{Map,Value}` interfaces further down into
the deployment engine so that configuration can be decrypted nearer to
its use.
This is the first part of the fix for pulumi/pulumi-ppc#112.
As documented in issue #616, the inputs/defaults/outputs model we have
today has fundamental problems. The crux of the issue is that our
current design requires that defaults present in the old state of a
resource are applied to the new inputs for that resource.
Unfortunately, it is not possible for the engine to decide which
defaults remain applicable and which do not; only the provider has that
knowledge.
These changes take a more tactical approach to resolving this issue than
that originally proposed in #616 that avoids breaking compatibility with
existing checkpoints. Rather than treating the Pulumi inputs as the
provider input properties for a resource, these inputs are first
translated by `Check`. In order to accommodate provider defaults that
were chosen for the old resource but should not change for the new,
`Check` now takes the old provider inputs as well as the new Pulumi
inputs. Rather than the Pulumi inputs and provider defaults, the
provider inputs returned by `Check` are recorded in the checkpoint file.
Put simply, these changes remove defaults as a first-class concept
(except inasmuch as is required to retain the ability to read old
checkpoint files) and move the responsibilty for manging and
merging defaults into the provider that supplies them.
Fixes#616.
This change adds a new manifest section to the checkpoint files.
The existing time moves into it, and we add to it the version of
the Pulumi CLI that created it, along with the names, types, and
versions of all plugins used to generate the file. There is a
magic cookie that we also use during verification.
This is to help keep us sane when debugging problems "in the wild,"
and I'm sure we will add more to it over time (checksum, etc).
For example, after an up, you can now see this in `pulumi stack`:
```
Current stack is demo:
Last updated at 2017-12-01 13:48:49.815740523 -0800 PST
Pulumi version v0.8.3-79-g1ab99ad
Plugin pulumi-provider-aws [resource] version v0.8.3-22-g4363e77
Plugin pulumi-langhost-nodejs [language] version v0.8.3-79-g77bb6b6
Checkpoint file is /Users/joeduffy/dev/code/src/github.com/pulumi/pulumi-aws/.pulumi/stacks/webserver/demo.json
```
This addresses pulumi/pulumi#628.
The prior change was incorrectly handling snapshotting of replacement
operations. Further, in hindsight, the older model of having steps
manage their interaction with the snapshot marking was clearer, so
I've essentially brought that back, merging it with the other changes.
This change simplifies the necessary RPC changes for components.
Instead of a Begin/End pair, which complicates the whole system
because now we have the opportunity of a missing End call, we will
simply let RPCs come in that append outputs to existing states.
We need to invoke the post-step event hook *after* updating the
state snapshots, so that it will write out the updated state.
We also need to re-serialize the snapshot again after we receive
updated output properties, otherwise they could be missing if this
happens to be the last resource (e.g., as in Stacks).
This change brings back component outputs to the overall system again.
In doing so, it generally overhauls the way we do resource RPCs a bit:
* Instead of RegisterResource and CompleteResource, we call these
BeginRegisterResource and EndRegisterResource, which begins to model
these as effectively "asynchronous" resource requests. This should also
help with parallelism (https://github.com/pulumi/pulumi/issues/106).
* Flip the CLI/engine a little on its head. Rather than it driving the
planning and deployment process, we move more to a model where it
simply observes it. This is done by implementing an event handler
interface with three events: OnResourceStepPre, OnResourceStepPost,
and OnResourceComplete. The first two are invoked immediately before
and after any step operation, and the latter is invoked whenever a
EndRegisterResource comes in. The reason for the asymmetry here is
that the checkpointing logic in the deployment engine is largely
untouched (intentionally, as this is a sensitive part of the system),
and so the "begin"/"end" nature doesn't flow through faithfully.
* Also make the engine more event-oriented in its terminology and the
way it handles the incoming BeginRegisterResource and
EndRegisterResource events from the language host. This is the first
step down a long road of incrementally refactoring the engine to work
this way, a necessary prerequisite for parallelism.
This change switches from child lists to parent pointers, in the
way resource ancestries are represented. This cleans up a fair bit
of the old parenting logic, including all notion of ambient parent
scopes (and will notably address pulumi/pulumi#435).
This lets us show a more parent/child display in the output when
doing planning and updating. For instance, here is an update of
a lambda's text, which is logically part of a cloud timer:
* cloud:timer:Timer: (same)
[urn=urn:pulumi:malta::lm-cloud:☁️timer:Timer::lm-cts-malta-job-CleanSnapshots]
* cloud:function:Function: (same)
[urn=urn:pulumi:malta::lm-cloud:☁️function:Function::lm-cts-malta-job-CleanSnapshots]
* aws:serverless:Function: (same)
[urn=urn:pulumi:malta::lm-cloud::aws:serverless:Function::lm-cts-malta-job-CleanSnapshots]
~ aws:lambda/function:Function: (modify)
[id=lm-cts-malta-job-CleanSnapshots-fee4f3bf41280741]
[urn=urn:pulumi:malta::lm-cloud::aws:lambda/function:Function::lm-cts-malta-job-CleanSnapshots]
- code : archive(assets:2092f44) {
// etc etc etc
Note that we still get walls of text, but this will be actually
quite nice when combined with pulumi/pulumi#454.
I've also suppressed printing properties that didn't change during
updates when --detailed was not passed, and also suppressed empty
strings and zero-length arrays (since TF uses these as defaults in
many places and it just makes creation and deletion quite verbose).
Note that this is a far cry from everything we can possibly do
here as part of pulumi/pulumi#340 (and even pulumi/pulumi#417).
But it's a good start towards taming some of our output spew.
This change adds back component output properties. Doing so
requires splitting the RPC interface for creating resources in
half, with an initial RegisterResource which contains all of the
input properties, and a final CompleteResource which optionally
contains any output properties synthesized by the component.
This change switches from child lists to parent pointers, in the
way resource ancestries are represented. This cleans up a fair bit
of the old parenting logic, including all notion of ambient parent
scopes (and will notably address pulumi/pulumi#435).
This lets us show a more parent/child display in the output when
doing planning and updating. For instance, here is an update of
a lambda's text, which is logically part of a cloud timer:
* cloud:timer:Timer: (same)
[urn=urn:pulumi:malta::lm-cloud:☁️timer:Timer::lm-cts-malta-job-CleanSnapshots]
* cloud:function:Function: (same)
[urn=urn:pulumi:malta::lm-cloud:☁️function:Function::lm-cts-malta-job-CleanSnapshots]
* aws:serverless:Function: (same)
[urn=urn:pulumi:malta::lm-cloud::aws:serverless:Function::lm-cts-malta-job-CleanSnapshots]
~ aws:lambda/function:Function: (modify)
[id=lm-cts-malta-job-CleanSnapshots-fee4f3bf41280741]
[urn=urn:pulumi:malta::lm-cloud::aws:lambda/function:Function::lm-cts-malta-job-CleanSnapshots]
- code : archive(assets:2092f44) {
// etc etc etc
Note that we still get walls of text, but this will be actually
quite nice when combined with pulumi/pulumi#454.
I've also suppressed printing properties that didn't change during
updates when --detailed was not passed, and also suppressed empty
strings and zero-length arrays (since TF uses these as defaults in
many places and it just makes creation and deletion quite verbose).
Note that this is a far cry from everything we can possibly do
here as part of pulumi/pulumi#340 (and even pulumi/pulumi#417).
But it's a good start towards taming some of our output spew.
When producing a snapshot for a plan, we have two resource DAGs. One of
these is the base DAG for the plan; the other is the current DAG for the
plan. Any resource r may be present in both DAGs. In order to produce a
snapshot, we need to merge these DAGs such that all resource
dependencies are correctly preserved. Conceptually, the merge proceeds
as follows:
- Begin with an empty merged DAG.
- For each resource r in the current DAG, insert r and its outgoing
edges into the merged DAG.
- For each resource r in the base DAG:
- If r is in the merged DAG, we are done: if the resource is in the
merged DAG, it must have been in the current DAG, which accurately
captures its current dependencies.
- If r is not in the merged DAG, insert it and its outgoing edges
into the merged DAG.
Physically, however, each DAG is represented as list of resources
without explicit dependency edges. In place of edges, it is assumed that
the list represents a valid topological sort of its source DAG. Thus,
any resource r at index i in a list L must be assumed to be dependent on
all resources in L with index j s.t. j < i. Due to this representation,
we implement the algorithm above as follows to produce a merged list
that represents a valid topological sort of the merged DAG:
- Begin with an empty merged list.
- For each resource r in the current list, append r to the merged list.
r must be in a correct location in the merged list, as its position
relative to its assumed dependencies has not changed.
- For each resource r in the base list:
- If r is in the merged list, we are done by the logic given in the
original algorithm.
- If r is not in the merged list, append r to the merged list. r
must be in a correct location in the merged list:
- If any of r's dependencies were in the current list, they must
already be in the merged list and their relative order w.r.t.
r has not changed.
- If any of r's dependencies were not in the current list, they
must already be in the merged list, as they would have been
appended to the list before r.
Prior to these changes, we had been performing these operations in
reverse order: we would start by appending any resources in the old list
that were not in the new list, then append the whole of the new list.
This caused out-of-order resources when a program that produced pending
deletions failed to run to completion.
Fixes#572.
We were previously calling configure on each package once per time it was mentioned in the config. We only need to call it once ever as we pass the full bag of relevent config through on that one call.
