pulumi/pkg/resource/snapshot.go

147 lines
6.2 KiB
Go
Raw Normal View History

2017-02-25 16:25:33 +01:00
// Copyright 2016 Pulumi, Inc. All rights reserved.
Begin resource modeling and planning This change introduces a new package, pkg/resource, that will form the foundation for actually performing deployment plans and applications. It contains the following key abstractions: * resource.Provider is a wrapper around the CRUD operations exposed by underlying resource plugins. It will eventually defer to resource.Plugin, which itself defers -- over an RPC interface -- to the actual plugin, one per package exposing resources. The provider will also understand how to load, cache, and overall manage the lifetime of each plugin. * resource.Resource is the actual resource object. This is created from the overall evaluation object graph, but is simplified. It contains only serializable properties, for example. Inter-resource references are translated into serializable monikers as part of creating the resource. * resource.Moniker is a serializable string that uniquely identifies a resource in the Mu system. This is in contrast to resource IDs, which are generated by resource providers and generally opaque to the Mu system. See marapongo/mu#69 for more information about monikers and some of their challenges (namely, designing a stable algorithm). * resource.Snapshot is a "snapshot" taken from a graph of resources. This is a transitive closure of state representing one possible configuration of a given environment. This is what plans are created from. Eventually, two snapshots will be diffable, in order to perform incremental updates. One way of thinking about this is that a snapshot of the old world's state is advanced, one step at a time, until it reaches a desired snapshot of the new world's state. * resource.Plan is a plan for carrying out desired CRUD operations on a target environment. Each plan consists of zero-to-many Steps, each of which has a CRUD operation type, a resource target, and a next step. This is an enumerator because it is possible the plan will evolve -- and introduce new steps -- as it is carried out (hence, the Next() method). At the moment, this is linearized; eventually, we want to make this more "graph-like" so that we can exploit available parallelism within the dependencies. There are tons of TODOs remaining. However, the `mu plan` command is functioning with these new changes -- including colorization FTW -- so I'm landing it now. This is part of marapongo/mu#38 and marapongo/mu#41.
2017-02-17 21:31:48 +01:00
package resource
import (
Redo object monikers This change overhauls the way we do object monikers. The old mechanism, generating monikers using graph paths, was far too brittle and prone to collisions. The new approach mixes some amount of "automatic scoping" plus some "explicit naming." Although there is some explicitness, this is arguably a good thing, as the monikers will be relatable back to the source more readily by developers inspecting the graph and resource state. Each moniker has four parts: <Namespace>::<AllocModule>::<Type>::<Name> wherein each element is the following: <Namespace> The namespace being deployed into <AllocModule> The module in which the object was allocated <Type> The type of the resource <Name> The assigned name of the resource The <Namespace> is essentially the deployment target -- so "prod", "stage", etc -- although it is more general purpose to allow for future namespacing within a target (e.g., "prod/customer1", etc); for now this is rudimentary, however, see marapongo/mu#94. The <AllocModule> is the token for the code that contained the 'new' that led to this object being created. In the future, we may wish to extend this to also track the module under evaluation. (This is a nice aspect of monikers; they can become arbitrarily complex, so long as they are precise, and not prone to false positives/negatives.) The <Name> warrants more discussion. The resource provider is consulted via a new gRPC method, Name, that fetches the name. How the provider does this is entirely up to it. For some resource types, the resource may have properties that developers must set (e.g., `new Bucket("foo")`); for other providers, perhaps the resource intrinsically has a property that explicitly and uniquely qualifies the object (e.g., AWS SecurityGroups, via `new SecurityGroup({groupName: "my-sg"}`); and finally, it's conceivable that a provider might auto-generate the name (e.g., such as an AWS Lambda whose name could simply be a hash of the source code contents). This should overall produce better results with respect to moniker collisions, ability to match resources, and the usability of the system.
2017-02-24 23:50:02 +01:00
"github.com/golang/glog"
2017-02-25 16:25:33 +01:00
"github.com/pulumi/coconut/pkg/compiler/core"
"github.com/pulumi/coconut/pkg/compiler/errors"
"github.com/pulumi/coconut/pkg/eval/heapstate"
"github.com/pulumi/coconut/pkg/eval/rt"
"github.com/pulumi/coconut/pkg/graph"
"github.com/pulumi/coconut/pkg/tokens"
"github.com/pulumi/coconut/pkg/util/contract"
Begin resource modeling and planning This change introduces a new package, pkg/resource, that will form the foundation for actually performing deployment plans and applications. It contains the following key abstractions: * resource.Provider is a wrapper around the CRUD operations exposed by underlying resource plugins. It will eventually defer to resource.Plugin, which itself defers -- over an RPC interface -- to the actual plugin, one per package exposing resources. The provider will also understand how to load, cache, and overall manage the lifetime of each plugin. * resource.Resource is the actual resource object. This is created from the overall evaluation object graph, but is simplified. It contains only serializable properties, for example. Inter-resource references are translated into serializable monikers as part of creating the resource. * resource.Moniker is a serializable string that uniquely identifies a resource in the Mu system. This is in contrast to resource IDs, which are generated by resource providers and generally opaque to the Mu system. See marapongo/mu#69 for more information about monikers and some of their challenges (namely, designing a stable algorithm). * resource.Snapshot is a "snapshot" taken from a graph of resources. This is a transitive closure of state representing one possible configuration of a given environment. This is what plans are created from. Eventually, two snapshots will be diffable, in order to perform incremental updates. One way of thinking about this is that a snapshot of the old world's state is advanced, one step at a time, until it reaches a desired snapshot of the new world's state. * resource.Plan is a plan for carrying out desired CRUD operations on a target environment. Each plan consists of zero-to-many Steps, each of which has a CRUD operation type, a resource target, and a next step. This is an enumerator because it is possible the plan will evolve -- and introduce new steps -- as it is carried out (hence, the Next() method). At the moment, this is linearized; eventually, we want to make this more "graph-like" so that we can exploit available parallelism within the dependencies. There are tons of TODOs remaining. However, the `mu plan` command is functioning with these new changes -- including colorization FTW -- so I'm landing it now. This is part of marapongo/mu#38 and marapongo/mu#41.
2017-02-17 21:31:48 +01:00
)
// Snapshot is a view of a collection of resources in an environment at a point in time. It describes resources; their
// IDs, names, and properties; their dependencies; and more. A snapshot is a diffable entity and can be used to create
// or apply an infrastructure deployment plan in order to make reality match the snapshot state.
type Snapshot interface {
Ctx() *Context // fetches the context for this snapshot.
Namespace() tokens.QName // the namespace target being deployed into.
Pkg() tokens.Package // the package from which this snapshot came.
Args() core.Args // the arguments used to compile this package.
Resources() []Resource // a topologically sorted list of resources (based on dependencies).
ResourceByID(id ID, t tokens.Type) Resource // looks up a resource by ID and type.
ResourceByURN(urn URN) Resource // looks up a resource by its URN.
ResourceByObject(obj *rt.Object) Resource // looks up a resource by its object.
Begin resource modeling and planning This change introduces a new package, pkg/resource, that will form the foundation for actually performing deployment plans and applications. It contains the following key abstractions: * resource.Provider is a wrapper around the CRUD operations exposed by underlying resource plugins. It will eventually defer to resource.Plugin, which itself defers -- over an RPC interface -- to the actual plugin, one per package exposing resources. The provider will also understand how to load, cache, and overall manage the lifetime of each plugin. * resource.Resource is the actual resource object. This is created from the overall evaluation object graph, but is simplified. It contains only serializable properties, for example. Inter-resource references are translated into serializable monikers as part of creating the resource. * resource.Moniker is a serializable string that uniquely identifies a resource in the Mu system. This is in contrast to resource IDs, which are generated by resource providers and generally opaque to the Mu system. See marapongo/mu#69 for more information about monikers and some of their challenges (namely, designing a stable algorithm). * resource.Snapshot is a "snapshot" taken from a graph of resources. This is a transitive closure of state representing one possible configuration of a given environment. This is what plans are created from. Eventually, two snapshots will be diffable, in order to perform incremental updates. One way of thinking about this is that a snapshot of the old world's state is advanced, one step at a time, until it reaches a desired snapshot of the new world's state. * resource.Plan is a plan for carrying out desired CRUD operations on a target environment. Each plan consists of zero-to-many Steps, each of which has a CRUD operation type, a resource target, and a next step. This is an enumerator because it is possible the plan will evolve -- and introduce new steps -- as it is carried out (hence, the Next() method). At the moment, this is linearized; eventually, we want to make this more "graph-like" so that we can exploit available parallelism within the dependencies. There are tons of TODOs remaining. However, the `mu plan` command is functioning with these new changes -- including colorization FTW -- so I'm landing it now. This is part of marapongo/mu#38 and marapongo/mu#41.
2017-02-17 21:31:48 +01:00
}
// NewSnapshot creates a snapshot from the given arguments. The resources must be in topologically sorted order.
func NewSnapshot(ctx *Context, ns tokens.QName, pkg tokens.Package,
args core.Args, resources []Resource) Snapshot {
return &snapshot{ctx, ns, pkg, args, resources}
}
// NewGraphSnapshot takes an object graph and produces a resource snapshot from it. It understands how to name
// resources based on their position within the graph and how to identify and record dependencies. This function can
// fail dynamically if the input graph did not satisfy the preconditions for resource graphs (like that it is a DAG).
func NewGraphSnapshot(ctx *Context, ns tokens.QName, pkg tokens.Package, args core.Args,
heap *heapstate.Heap, old Snapshot) (Snapshot, error) {
// If the old snapshot is non-nil, we need to register old IDs so they will be found below.
if old != nil {
for _, res := range old.Resources() {
contract.Assert(res.HasID())
ctx.URNOldIDs[res.URN()] = res.ID()
}
}
Redo object monikers This change overhauls the way we do object monikers. The old mechanism, generating monikers using graph paths, was far too brittle and prone to collisions. The new approach mixes some amount of "automatic scoping" plus some "explicit naming." Although there is some explicitness, this is arguably a good thing, as the monikers will be relatable back to the source more readily by developers inspecting the graph and resource state. Each moniker has four parts: <Namespace>::<AllocModule>::<Type>::<Name> wherein each element is the following: <Namespace> The namespace being deployed into <AllocModule> The module in which the object was allocated <Type> The type of the resource <Name> The assigned name of the resource The <Namespace> is essentially the deployment target -- so "prod", "stage", etc -- although it is more general purpose to allow for future namespacing within a target (e.g., "prod/customer1", etc); for now this is rudimentary, however, see marapongo/mu#94. The <AllocModule> is the token for the code that contained the 'new' that led to this object being created. In the future, we may wish to extend this to also track the module under evaluation. (This is a nice aspect of monikers; they can become arbitrarily complex, so long as they are precise, and not prone to false positives/negatives.) The <Name> warrants more discussion. The resource provider is consulted via a new gRPC method, Name, that fetches the name. How the provider does this is entirely up to it. For some resource types, the resource may have properties that developers must set (e.g., `new Bucket("foo")`); for other providers, perhaps the resource intrinsically has a property that explicitly and uniquely qualifies the object (e.g., AWS SecurityGroups, via `new SecurityGroup({groupName: "my-sg"}`); and finally, it's conceivable that a provider might auto-generate the name (e.g., such as an AWS Lambda whose name could simply be a hash of the source code contents). This should overall produce better results with respect to moniker collisions, ability to match resources, and the usability of the system.
2017-02-24 23:50:02 +01:00
// Topologically sort the entire heapstate (in dependency order) and extract just the resource objects.
resobjs, err := topsort(ctx, heap.G)
if err != nil {
return nil, err
}
// Next, name all resources, create their URNs and objects, and maps that we will use. Note that we must do
// this in DAG order (guaranteed by our topological sort above), so that referenced URNs are available.
resources, err := createResources(ctx, ns, heap, resobjs)
Begin resource modeling and planning This change introduces a new package, pkg/resource, that will form the foundation for actually performing deployment plans and applications. It contains the following key abstractions: * resource.Provider is a wrapper around the CRUD operations exposed by underlying resource plugins. It will eventually defer to resource.Plugin, which itself defers -- over an RPC interface -- to the actual plugin, one per package exposing resources. The provider will also understand how to load, cache, and overall manage the lifetime of each plugin. * resource.Resource is the actual resource object. This is created from the overall evaluation object graph, but is simplified. It contains only serializable properties, for example. Inter-resource references are translated into serializable monikers as part of creating the resource. * resource.Moniker is a serializable string that uniquely identifies a resource in the Mu system. This is in contrast to resource IDs, which are generated by resource providers and generally opaque to the Mu system. See marapongo/mu#69 for more information about monikers and some of their challenges (namely, designing a stable algorithm). * resource.Snapshot is a "snapshot" taken from a graph of resources. This is a transitive closure of state representing one possible configuration of a given environment. This is what plans are created from. Eventually, two snapshots will be diffable, in order to perform incremental updates. One way of thinking about this is that a snapshot of the old world's state is advanced, one step at a time, until it reaches a desired snapshot of the new world's state. * resource.Plan is a plan for carrying out desired CRUD operations on a target environment. Each plan consists of zero-to-many Steps, each of which has a CRUD operation type, a resource target, and a next step. This is an enumerator because it is possible the plan will evolve -- and introduce new steps -- as it is carried out (hence, the Next() method). At the moment, this is linearized; eventually, we want to make this more "graph-like" so that we can exploit available parallelism within the dependencies. There are tons of TODOs remaining. However, the `mu plan` command is functioning with these new changes -- including colorization FTW -- so I'm landing it now. This is part of marapongo/mu#38 and marapongo/mu#41.
2017-02-17 21:31:48 +01:00
if err != nil {
return nil, err
}
return NewSnapshot(ctx, ns, pkg, args, resources), nil
Begin resource modeling and planning This change introduces a new package, pkg/resource, that will form the foundation for actually performing deployment plans and applications. It contains the following key abstractions: * resource.Provider is a wrapper around the CRUD operations exposed by underlying resource plugins. It will eventually defer to resource.Plugin, which itself defers -- over an RPC interface -- to the actual plugin, one per package exposing resources. The provider will also understand how to load, cache, and overall manage the lifetime of each plugin. * resource.Resource is the actual resource object. This is created from the overall evaluation object graph, but is simplified. It contains only serializable properties, for example. Inter-resource references are translated into serializable monikers as part of creating the resource. * resource.Moniker is a serializable string that uniquely identifies a resource in the Mu system. This is in contrast to resource IDs, which are generated by resource providers and generally opaque to the Mu system. See marapongo/mu#69 for more information about monikers and some of their challenges (namely, designing a stable algorithm). * resource.Snapshot is a "snapshot" taken from a graph of resources. This is a transitive closure of state representing one possible configuration of a given environment. This is what plans are created from. Eventually, two snapshots will be diffable, in order to perform incremental updates. One way of thinking about this is that a snapshot of the old world's state is advanced, one step at a time, until it reaches a desired snapshot of the new world's state. * resource.Plan is a plan for carrying out desired CRUD operations on a target environment. Each plan consists of zero-to-many Steps, each of which has a CRUD operation type, a resource target, and a next step. This is an enumerator because it is possible the plan will evolve -- and introduce new steps -- as it is carried out (hence, the Next() method). At the moment, this is linearized; eventually, we want to make this more "graph-like" so that we can exploit available parallelism within the dependencies. There are tons of TODOs remaining. However, the `mu plan` command is functioning with these new changes -- including colorization FTW -- so I'm landing it now. This is part of marapongo/mu#38 and marapongo/mu#41.
2017-02-17 21:31:48 +01:00
}
type snapshot struct {
ctx *Context // the context shared by all operations in this snapshot.
ns tokens.QName // the namespace target being deployed into.
pkg tokens.Package // the package from which this snapshot came.
args core.Args // the arguments used to compile this package.
resources []Resource // the topologically sorted linearized list of resources.
}
func (s *snapshot) Ctx() *Context { return s.ctx }
func (s *snapshot) Namespace() tokens.QName { return s.ns }
func (s *snapshot) Pkg() tokens.Package { return s.pkg }
func (s *snapshot) Args() core.Args { return s.args }
func (s *snapshot) Resources() []Resource { return s.resources }
func (s *snapshot) ResourceByID(id ID, t tokens.Type) Resource {
contract.Failf("TODO: not yet implemented")
return nil
Begin resource modeling and planning This change introduces a new package, pkg/resource, that will form the foundation for actually performing deployment plans and applications. It contains the following key abstractions: * resource.Provider is a wrapper around the CRUD operations exposed by underlying resource plugins. It will eventually defer to resource.Plugin, which itself defers -- over an RPC interface -- to the actual plugin, one per package exposing resources. The provider will also understand how to load, cache, and overall manage the lifetime of each plugin. * resource.Resource is the actual resource object. This is created from the overall evaluation object graph, but is simplified. It contains only serializable properties, for example. Inter-resource references are translated into serializable monikers as part of creating the resource. * resource.Moniker is a serializable string that uniquely identifies a resource in the Mu system. This is in contrast to resource IDs, which are generated by resource providers and generally opaque to the Mu system. See marapongo/mu#69 for more information about monikers and some of their challenges (namely, designing a stable algorithm). * resource.Snapshot is a "snapshot" taken from a graph of resources. This is a transitive closure of state representing one possible configuration of a given environment. This is what plans are created from. Eventually, two snapshots will be diffable, in order to perform incremental updates. One way of thinking about this is that a snapshot of the old world's state is advanced, one step at a time, until it reaches a desired snapshot of the new world's state. * resource.Plan is a plan for carrying out desired CRUD operations on a target environment. Each plan consists of zero-to-many Steps, each of which has a CRUD operation type, a resource target, and a next step. This is an enumerator because it is possible the plan will evolve -- and introduce new steps -- as it is carried out (hence, the Next() method). At the moment, this is linearized; eventually, we want to make this more "graph-like" so that we can exploit available parallelism within the dependencies. There are tons of TODOs remaining. However, the `mu plan` command is functioning with these new changes -- including colorization FTW -- so I'm landing it now. This is part of marapongo/mu#38 and marapongo/mu#41.
2017-02-17 21:31:48 +01:00
}
func (s *snapshot) ResourceByURN(urn URN) Resource { return s.ctx.URNRes[urn] }
Redo object monikers This change overhauls the way we do object monikers. The old mechanism, generating monikers using graph paths, was far too brittle and prone to collisions. The new approach mixes some amount of "automatic scoping" plus some "explicit naming." Although there is some explicitness, this is arguably a good thing, as the monikers will be relatable back to the source more readily by developers inspecting the graph and resource state. Each moniker has four parts: <Namespace>::<AllocModule>::<Type>::<Name> wherein each element is the following: <Namespace> The namespace being deployed into <AllocModule> The module in which the object was allocated <Type> The type of the resource <Name> The assigned name of the resource The <Namespace> is essentially the deployment target -- so "prod", "stage", etc -- although it is more general purpose to allow for future namespacing within a target (e.g., "prod/customer1", etc); for now this is rudimentary, however, see marapongo/mu#94. The <AllocModule> is the token for the code that contained the 'new' that led to this object being created. In the future, we may wish to extend this to also track the module under evaluation. (This is a nice aspect of monikers; they can become arbitrarily complex, so long as they are precise, and not prone to false positives/negatives.) The <Name> warrants more discussion. The resource provider is consulted via a new gRPC method, Name, that fetches the name. How the provider does this is entirely up to it. For some resource types, the resource may have properties that developers must set (e.g., `new Bucket("foo")`); for other providers, perhaps the resource intrinsically has a property that explicitly and uniquely qualifies the object (e.g., AWS SecurityGroups, via `new SecurityGroup({groupName: "my-sg"}`); and finally, it's conceivable that a provider might auto-generate the name (e.g., such as an AWS Lambda whose name could simply be a hash of the source code contents). This should overall produce better results with respect to moniker collisions, ability to match resources, and the usability of the system.
2017-02-24 23:50:02 +01:00
func (s *snapshot) ResourceByObject(obj *rt.Object) Resource { return s.ctx.ObjRes[obj] }
// createResources uses a graph to create URNs and resource objects for every resource within. It
// returns two maps for further use: a map of vertex to its new resource object, and a map of vertex to its URN.
func createResources(ctx *Context, ns tokens.QName, heap *heapstate.Heap, resobjs []*rt.Object) ([]Resource, error) {
Redo object monikers This change overhauls the way we do object monikers. The old mechanism, generating monikers using graph paths, was far too brittle and prone to collisions. The new approach mixes some amount of "automatic scoping" plus some "explicit naming." Although there is some explicitness, this is arguably a good thing, as the monikers will be relatable back to the source more readily by developers inspecting the graph and resource state. Each moniker has four parts: <Namespace>::<AllocModule>::<Type>::<Name> wherein each element is the following: <Namespace> The namespace being deployed into <AllocModule> The module in which the object was allocated <Type> The type of the resource <Name> The assigned name of the resource The <Namespace> is essentially the deployment target -- so "prod", "stage", etc -- although it is more general purpose to allow for future namespacing within a target (e.g., "prod/customer1", etc); for now this is rudimentary, however, see marapongo/mu#94. The <AllocModule> is the token for the code that contained the 'new' that led to this object being created. In the future, we may wish to extend this to also track the module under evaluation. (This is a nice aspect of monikers; they can become arbitrarily complex, so long as they are precise, and not prone to false positives/negatives.) The <Name> warrants more discussion. The resource provider is consulted via a new gRPC method, Name, that fetches the name. How the provider does this is entirely up to it. For some resource types, the resource may have properties that developers must set (e.g., `new Bucket("foo")`); for other providers, perhaps the resource intrinsically has a property that explicitly and uniquely qualifies the object (e.g., AWS SecurityGroups, via `new SecurityGroup({groupName: "my-sg"}`); and finally, it's conceivable that a provider might auto-generate the name (e.g., such as an AWS Lambda whose name could simply be a hash of the source code contents). This should overall produce better results with respect to moniker collisions, ability to match resources, and the usability of the system.
2017-02-24 23:50:02 +01:00
var resources []Resource
for _, resobj := range resobjs {
// Create an object resource without a URN.
Redo object monikers This change overhauls the way we do object monikers. The old mechanism, generating monikers using graph paths, was far too brittle and prone to collisions. The new approach mixes some amount of "automatic scoping" plus some "explicit naming." Although there is some explicitness, this is arguably a good thing, as the monikers will be relatable back to the source more readily by developers inspecting the graph and resource state. Each moniker has four parts: <Namespace>::<AllocModule>::<Type>::<Name> wherein each element is the following: <Namespace> The namespace being deployed into <AllocModule> The module in which the object was allocated <Type> The type of the resource <Name> The assigned name of the resource The <Namespace> is essentially the deployment target -- so "prod", "stage", etc -- although it is more general purpose to allow for future namespacing within a target (e.g., "prod/customer1", etc); for now this is rudimentary, however, see marapongo/mu#94. The <AllocModule> is the token for the code that contained the 'new' that led to this object being created. In the future, we may wish to extend this to also track the module under evaluation. (This is a nice aspect of monikers; they can become arbitrarily complex, so long as they are precise, and not prone to false positives/negatives.) The <Name> warrants more discussion. The resource provider is consulted via a new gRPC method, Name, that fetches the name. How the provider does this is entirely up to it. For some resource types, the resource may have properties that developers must set (e.g., `new Bucket("foo")`); for other providers, perhaps the resource intrinsically has a property that explicitly and uniquely qualifies the object (e.g., AWS SecurityGroups, via `new SecurityGroup({groupName: "my-sg"}`); and finally, it's conceivable that a provider might auto-generate the name (e.g., such as an AWS Lambda whose name could simply be a hash of the source code contents). This should overall produce better results with respect to moniker collisions, ability to match resources, and the usability of the system.
2017-02-24 23:50:02 +01:00
res := NewObjectResource(ctx, resobj)
// Now fetch this resource's name by looking up its provider and doing an RPC.
t := resobj.Type().TypeToken()
prov, err := ctx.Provider(t.Package())
if err != nil {
return nil, err
}
Redo object monikers This change overhauls the way we do object monikers. The old mechanism, generating monikers using graph paths, was far too brittle and prone to collisions. The new approach mixes some amount of "automatic scoping" plus some "explicit naming." Although there is some explicitness, this is arguably a good thing, as the monikers will be relatable back to the source more readily by developers inspecting the graph and resource state. Each moniker has four parts: <Namespace>::<AllocModule>::<Type>::<Name> wherein each element is the following: <Namespace> The namespace being deployed into <AllocModule> The module in which the object was allocated <Type> The type of the resource <Name> The assigned name of the resource The <Namespace> is essentially the deployment target -- so "prod", "stage", etc -- although it is more general purpose to allow for future namespacing within a target (e.g., "prod/customer1", etc); for now this is rudimentary, however, see marapongo/mu#94. The <AllocModule> is the token for the code that contained the 'new' that led to this object being created. In the future, we may wish to extend this to also track the module under evaluation. (This is a nice aspect of monikers; they can become arbitrarily complex, so long as they are precise, and not prone to false positives/negatives.) The <Name> warrants more discussion. The resource provider is consulted via a new gRPC method, Name, that fetches the name. How the provider does this is entirely up to it. For some resource types, the resource may have properties that developers must set (e.g., `new Bucket("foo")`); for other providers, perhaps the resource intrinsically has a property that explicitly and uniquely qualifies the object (e.g., AWS SecurityGroups, via `new SecurityGroup({groupName: "my-sg"}`); and finally, it's conceivable that a provider might auto-generate the name (e.g., such as an AWS Lambda whose name could simply be a hash of the source code contents). This should overall produce better results with respect to moniker collisions, ability to match resources, and the usability of the system.
2017-02-24 23:50:02 +01:00
name, err := prov.Name(t, res.Properties())
if err != nil {
return nil, err
}
// Now compute a unique URN for this object and ensure we haven't had any collisions.
alloc := heap.Alloc(resobj)
urn := NewURN(ns, alloc.Mod.Tok, t, name)
glog.V(7).Infof("Resource URN computed: %v", urn)
if _, exists := ctx.URNRes[urn]; exists {
// If this URN is already in use, issue an error, ignore this one, and break. The break is necessary
// because subsequent resources might contain references to this URN and would fail to find it.
ctx.Diag.Errorf(errors.ErrorDuplicateURNNames.At(alloc.Loc), urn)
break
} else {
res.SetURN(urn)
ctx.ObjRes[resobj] = res
ctx.URNRes[urn] = res
ctx.ObjURN[resobj] = urn
}
Redo object monikers This change overhauls the way we do object monikers. The old mechanism, generating monikers using graph paths, was far too brittle and prone to collisions. The new approach mixes some amount of "automatic scoping" plus some "explicit naming." Although there is some explicitness, this is arguably a good thing, as the monikers will be relatable back to the source more readily by developers inspecting the graph and resource state. Each moniker has four parts: <Namespace>::<AllocModule>::<Type>::<Name> wherein each element is the following: <Namespace> The namespace being deployed into <AllocModule> The module in which the object was allocated <Type> The type of the resource <Name> The assigned name of the resource The <Namespace> is essentially the deployment target -- so "prod", "stage", etc -- although it is more general purpose to allow for future namespacing within a target (e.g., "prod/customer1", etc); for now this is rudimentary, however, see marapongo/mu#94. The <AllocModule> is the token for the code that contained the 'new' that led to this object being created. In the future, we may wish to extend this to also track the module under evaluation. (This is a nice aspect of monikers; they can become arbitrarily complex, so long as they are precise, and not prone to false positives/negatives.) The <Name> warrants more discussion. The resource provider is consulted via a new gRPC method, Name, that fetches the name. How the provider does this is entirely up to it. For some resource types, the resource may have properties that developers must set (e.g., `new Bucket("foo")`); for other providers, perhaps the resource intrinsically has a property that explicitly and uniquely qualifies the object (e.g., AWS SecurityGroups, via `new SecurityGroup({groupName: "my-sg"}`); and finally, it's conceivable that a provider might auto-generate the name (e.g., such as an AWS Lambda whose name could simply be a hash of the source code contents). This should overall produce better results with respect to moniker collisions, ability to match resources, and the usability of the system.
2017-02-24 23:50:02 +01:00
resources = append(resources, res)
}
Redo object monikers This change overhauls the way we do object monikers. The old mechanism, generating monikers using graph paths, was far too brittle and prone to collisions. The new approach mixes some amount of "automatic scoping" plus some "explicit naming." Although there is some explicitness, this is arguably a good thing, as the monikers will be relatable back to the source more readily by developers inspecting the graph and resource state. Each moniker has four parts: <Namespace>::<AllocModule>::<Type>::<Name> wherein each element is the following: <Namespace> The namespace being deployed into <AllocModule> The module in which the object was allocated <Type> The type of the resource <Name> The assigned name of the resource The <Namespace> is essentially the deployment target -- so "prod", "stage", etc -- although it is more general purpose to allow for future namespacing within a target (e.g., "prod/customer1", etc); for now this is rudimentary, however, see marapongo/mu#94. The <AllocModule> is the token for the code that contained the 'new' that led to this object being created. In the future, we may wish to extend this to also track the module under evaluation. (This is a nice aspect of monikers; they can become arbitrarily complex, so long as they are precise, and not prone to false positives/negatives.) The <Name> warrants more discussion. The resource provider is consulted via a new gRPC method, Name, that fetches the name. How the provider does this is entirely up to it. For some resource types, the resource may have properties that developers must set (e.g., `new Bucket("foo")`); for other providers, perhaps the resource intrinsically has a property that explicitly and uniquely qualifies the object (e.g., AWS SecurityGroups, via `new SecurityGroup({groupName: "my-sg"}`); and finally, it's conceivable that a provider might auto-generate the name (e.g., such as an AWS Lambda whose name could simply be a hash of the source code contents). This should overall produce better results with respect to moniker collisions, ability to match resources, and the usability of the system.
2017-02-24 23:50:02 +01:00
return resources, nil
}
// topsort actually performs a topological sort on a resource graph.
Redo object monikers This change overhauls the way we do object monikers. The old mechanism, generating monikers using graph paths, was far too brittle and prone to collisions. The new approach mixes some amount of "automatic scoping" plus some "explicit naming." Although there is some explicitness, this is arguably a good thing, as the monikers will be relatable back to the source more readily by developers inspecting the graph and resource state. Each moniker has four parts: <Namespace>::<AllocModule>::<Type>::<Name> wherein each element is the following: <Namespace> The namespace being deployed into <AllocModule> The module in which the object was allocated <Type> The type of the resource <Name> The assigned name of the resource The <Namespace> is essentially the deployment target -- so "prod", "stage", etc -- although it is more general purpose to allow for future namespacing within a target (e.g., "prod/customer1", etc); for now this is rudimentary, however, see marapongo/mu#94. The <AllocModule> is the token for the code that contained the 'new' that led to this object being created. In the future, we may wish to extend this to also track the module under evaluation. (This is a nice aspect of monikers; they can become arbitrarily complex, so long as they are precise, and not prone to false positives/negatives.) The <Name> warrants more discussion. The resource provider is consulted via a new gRPC method, Name, that fetches the name. How the provider does this is entirely up to it. For some resource types, the resource may have properties that developers must set (e.g., `new Bucket("foo")`); for other providers, perhaps the resource intrinsically has a property that explicitly and uniquely qualifies the object (e.g., AWS SecurityGroups, via `new SecurityGroup({groupName: "my-sg"}`); and finally, it's conceivable that a provider might auto-generate the name (e.g., such as an AWS Lambda whose name could simply be a hash of the source code contents). This should overall produce better results with respect to moniker collisions, ability to match resources, and the usability of the system.
2017-02-24 23:50:02 +01:00
func topsort(ctx *Context, g graph.Graph) ([]*rt.Object, error) {
Begin resource modeling and planning This change introduces a new package, pkg/resource, that will form the foundation for actually performing deployment plans and applications. It contains the following key abstractions: * resource.Provider is a wrapper around the CRUD operations exposed by underlying resource plugins. It will eventually defer to resource.Plugin, which itself defers -- over an RPC interface -- to the actual plugin, one per package exposing resources. The provider will also understand how to load, cache, and overall manage the lifetime of each plugin. * resource.Resource is the actual resource object. This is created from the overall evaluation object graph, but is simplified. It contains only serializable properties, for example. Inter-resource references are translated into serializable monikers as part of creating the resource. * resource.Moniker is a serializable string that uniquely identifies a resource in the Mu system. This is in contrast to resource IDs, which are generated by resource providers and generally opaque to the Mu system. See marapongo/mu#69 for more information about monikers and some of their challenges (namely, designing a stable algorithm). * resource.Snapshot is a "snapshot" taken from a graph of resources. This is a transitive closure of state representing one possible configuration of a given environment. This is what plans are created from. Eventually, two snapshots will be diffable, in order to perform incremental updates. One way of thinking about this is that a snapshot of the old world's state is advanced, one step at a time, until it reaches a desired snapshot of the new world's state. * resource.Plan is a plan for carrying out desired CRUD operations on a target environment. Each plan consists of zero-to-many Steps, each of which has a CRUD operation type, a resource target, and a next step. This is an enumerator because it is possible the plan will evolve -- and introduce new steps -- as it is carried out (hence, the Next() method). At the moment, this is linearized; eventually, we want to make this more "graph-like" so that we can exploit available parallelism within the dependencies. There are tons of TODOs remaining. However, the `mu plan` command is functioning with these new changes -- including colorization FTW -- so I'm landing it now. This is part of marapongo/mu#38 and marapongo/mu#41.
2017-02-17 21:31:48 +01:00
// Sort the graph output so that it's a DAG; if it's got cycles, this can fail.
Redo object monikers This change overhauls the way we do object monikers. The old mechanism, generating monikers using graph paths, was far too brittle and prone to collisions. The new approach mixes some amount of "automatic scoping" plus some "explicit naming." Although there is some explicitness, this is arguably a good thing, as the monikers will be relatable back to the source more readily by developers inspecting the graph and resource state. Each moniker has four parts: <Namespace>::<AllocModule>::<Type>::<Name> wherein each element is the following: <Namespace> The namespace being deployed into <AllocModule> The module in which the object was allocated <Type> The type of the resource <Name> The assigned name of the resource The <Namespace> is essentially the deployment target -- so "prod", "stage", etc -- although it is more general purpose to allow for future namespacing within a target (e.g., "prod/customer1", etc); for now this is rudimentary, however, see marapongo/mu#94. The <AllocModule> is the token for the code that contained the 'new' that led to this object being created. In the future, we may wish to extend this to also track the module under evaluation. (This is a nice aspect of monikers; they can become arbitrarily complex, so long as they are precise, and not prone to false positives/negatives.) The <Name> warrants more discussion. The resource provider is consulted via a new gRPC method, Name, that fetches the name. How the provider does this is entirely up to it. For some resource types, the resource may have properties that developers must set (e.g., `new Bucket("foo")`); for other providers, perhaps the resource intrinsically has a property that explicitly and uniquely qualifies the object (e.g., AWS SecurityGroups, via `new SecurityGroup({groupName: "my-sg"}`); and finally, it's conceivable that a provider might auto-generate the name (e.g., such as an AWS Lambda whose name could simply be a hash of the source code contents). This should overall produce better results with respect to moniker collisions, ability to match resources, and the usability of the system.
2017-02-24 23:50:02 +01:00
// TODO: we want this to return a *graph*, not a linearized list, so that we can parallelize.
// TODO: it'd be nice to prune the graph to just the resource objects first, so we don't waste effort.
Begin resource modeling and planning This change introduces a new package, pkg/resource, that will form the foundation for actually performing deployment plans and applications. It contains the following key abstractions: * resource.Provider is a wrapper around the CRUD operations exposed by underlying resource plugins. It will eventually defer to resource.Plugin, which itself defers -- over an RPC interface -- to the actual plugin, one per package exposing resources. The provider will also understand how to load, cache, and overall manage the lifetime of each plugin. * resource.Resource is the actual resource object. This is created from the overall evaluation object graph, but is simplified. It contains only serializable properties, for example. Inter-resource references are translated into serializable monikers as part of creating the resource. * resource.Moniker is a serializable string that uniquely identifies a resource in the Mu system. This is in contrast to resource IDs, which are generated by resource providers and generally opaque to the Mu system. See marapongo/mu#69 for more information about monikers and some of their challenges (namely, designing a stable algorithm). * resource.Snapshot is a "snapshot" taken from a graph of resources. This is a transitive closure of state representing one possible configuration of a given environment. This is what plans are created from. Eventually, two snapshots will be diffable, in order to perform incremental updates. One way of thinking about this is that a snapshot of the old world's state is advanced, one step at a time, until it reaches a desired snapshot of the new world's state. * resource.Plan is a plan for carrying out desired CRUD operations on a target environment. Each plan consists of zero-to-many Steps, each of which has a CRUD operation type, a resource target, and a next step. This is an enumerator because it is possible the plan will evolve -- and introduce new steps -- as it is carried out (hence, the Next() method). At the moment, this is linearized; eventually, we want to make this more "graph-like" so that we can exploit available parallelism within the dependencies. There are tons of TODOs remaining. However, the `mu plan` command is functioning with these new changes -- including colorization FTW -- so I'm landing it now. This is part of marapongo/mu#38 and marapongo/mu#41.
2017-02-17 21:31:48 +01:00
sorted, err := graph.Topsort(g)
if err != nil {
Redo object monikers This change overhauls the way we do object monikers. The old mechanism, generating monikers using graph paths, was far too brittle and prone to collisions. The new approach mixes some amount of "automatic scoping" plus some "explicit naming." Although there is some explicitness, this is arguably a good thing, as the monikers will be relatable back to the source more readily by developers inspecting the graph and resource state. Each moniker has four parts: <Namespace>::<AllocModule>::<Type>::<Name> wherein each element is the following: <Namespace> The namespace being deployed into <AllocModule> The module in which the object was allocated <Type> The type of the resource <Name> The assigned name of the resource The <Namespace> is essentially the deployment target -- so "prod", "stage", etc -- although it is more general purpose to allow for future namespacing within a target (e.g., "prod/customer1", etc); for now this is rudimentary, however, see marapongo/mu#94. The <AllocModule> is the token for the code that contained the 'new' that led to this object being created. In the future, we may wish to extend this to also track the module under evaluation. (This is a nice aspect of monikers; they can become arbitrarily complex, so long as they are precise, and not prone to false positives/negatives.) The <Name> warrants more discussion. The resource provider is consulted via a new gRPC method, Name, that fetches the name. How the provider does this is entirely up to it. For some resource types, the resource may have properties that developers must set (e.g., `new Bucket("foo")`); for other providers, perhaps the resource intrinsically has a property that explicitly and uniquely qualifies the object (e.g., AWS SecurityGroups, via `new SecurityGroup({groupName: "my-sg"}`); and finally, it's conceivable that a provider might auto-generate the name (e.g., such as an AWS Lambda whose name could simply be a hash of the source code contents). This should overall produce better results with respect to moniker collisions, ability to match resources, and the usability of the system.
2017-02-24 23:50:02 +01:00
return nil, err
Begin resource modeling and planning This change introduces a new package, pkg/resource, that will form the foundation for actually performing deployment plans and applications. It contains the following key abstractions: * resource.Provider is a wrapper around the CRUD operations exposed by underlying resource plugins. It will eventually defer to resource.Plugin, which itself defers -- over an RPC interface -- to the actual plugin, one per package exposing resources. The provider will also understand how to load, cache, and overall manage the lifetime of each plugin. * resource.Resource is the actual resource object. This is created from the overall evaluation object graph, but is simplified. It contains only serializable properties, for example. Inter-resource references are translated into serializable monikers as part of creating the resource. * resource.Moniker is a serializable string that uniquely identifies a resource in the Mu system. This is in contrast to resource IDs, which are generated by resource providers and generally opaque to the Mu system. See marapongo/mu#69 for more information about monikers and some of their challenges (namely, designing a stable algorithm). * resource.Snapshot is a "snapshot" taken from a graph of resources. This is a transitive closure of state representing one possible configuration of a given environment. This is what plans are created from. Eventually, two snapshots will be diffable, in order to perform incremental updates. One way of thinking about this is that a snapshot of the old world's state is advanced, one step at a time, until it reaches a desired snapshot of the new world's state. * resource.Plan is a plan for carrying out desired CRUD operations on a target environment. Each plan consists of zero-to-many Steps, each of which has a CRUD operation type, a resource target, and a next step. This is an enumerator because it is possible the plan will evolve -- and introduce new steps -- as it is carried out (hence, the Next() method). At the moment, this is linearized; eventually, we want to make this more "graph-like" so that we can exploit available parallelism within the dependencies. There are tons of TODOs remaining. However, the `mu plan` command is functioning with these new changes -- including colorization FTW -- so I'm landing it now. This is part of marapongo/mu#38 and marapongo/mu#41.
2017-02-17 21:31:48 +01:00
}
// Now walk the list and prune out anything that isn't a resource.
Redo object monikers This change overhauls the way we do object monikers. The old mechanism, generating monikers using graph paths, was far too brittle and prone to collisions. The new approach mixes some amount of "automatic scoping" plus some "explicit naming." Although there is some explicitness, this is arguably a good thing, as the monikers will be relatable back to the source more readily by developers inspecting the graph and resource state. Each moniker has four parts: <Namespace>::<AllocModule>::<Type>::<Name> wherein each element is the following: <Namespace> The namespace being deployed into <AllocModule> The module in which the object was allocated <Type> The type of the resource <Name> The assigned name of the resource The <Namespace> is essentially the deployment target -- so "prod", "stage", etc -- although it is more general purpose to allow for future namespacing within a target (e.g., "prod/customer1", etc); for now this is rudimentary, however, see marapongo/mu#94. The <AllocModule> is the token for the code that contained the 'new' that led to this object being created. In the future, we may wish to extend this to also track the module under evaluation. (This is a nice aspect of monikers; they can become arbitrarily complex, so long as they are precise, and not prone to false positives/negatives.) The <Name> warrants more discussion. The resource provider is consulted via a new gRPC method, Name, that fetches the name. How the provider does this is entirely up to it. For some resource types, the resource may have properties that developers must set (e.g., `new Bucket("foo")`); for other providers, perhaps the resource intrinsically has a property that explicitly and uniquely qualifies the object (e.g., AWS SecurityGroups, via `new SecurityGroup({groupName: "my-sg"}`); and finally, it's conceivable that a provider might auto-generate the name (e.g., such as an AWS Lambda whose name could simply be a hash of the source code contents). This should overall produce better results with respect to moniker collisions, ability to match resources, and the usability of the system.
2017-02-24 23:50:02 +01:00
var resobjs []*rt.Object
Begin resource modeling and planning This change introduces a new package, pkg/resource, that will form the foundation for actually performing deployment plans and applications. It contains the following key abstractions: * resource.Provider is a wrapper around the CRUD operations exposed by underlying resource plugins. It will eventually defer to resource.Plugin, which itself defers -- over an RPC interface -- to the actual plugin, one per package exposing resources. The provider will also understand how to load, cache, and overall manage the lifetime of each plugin. * resource.Resource is the actual resource object. This is created from the overall evaluation object graph, but is simplified. It contains only serializable properties, for example. Inter-resource references are translated into serializable monikers as part of creating the resource. * resource.Moniker is a serializable string that uniquely identifies a resource in the Mu system. This is in contrast to resource IDs, which are generated by resource providers and generally opaque to the Mu system. See marapongo/mu#69 for more information about monikers and some of their challenges (namely, designing a stable algorithm). * resource.Snapshot is a "snapshot" taken from a graph of resources. This is a transitive closure of state representing one possible configuration of a given environment. This is what plans are created from. Eventually, two snapshots will be diffable, in order to perform incremental updates. One way of thinking about this is that a snapshot of the old world's state is advanced, one step at a time, until it reaches a desired snapshot of the new world's state. * resource.Plan is a plan for carrying out desired CRUD operations on a target environment. Each plan consists of zero-to-many Steps, each of which has a CRUD operation type, a resource target, and a next step. This is an enumerator because it is possible the plan will evolve -- and introduce new steps -- as it is carried out (hence, the Next() method). At the moment, this is linearized; eventually, we want to make this more "graph-like" so that we can exploit available parallelism within the dependencies. There are tons of TODOs remaining. However, the `mu plan` command is functioning with these new changes -- including colorization FTW -- so I'm landing it now. This is part of marapongo/mu#38 and marapongo/mu#41.
2017-02-17 21:31:48 +01:00
for _, v := range sorted {
Implement updates This change is a first whack at implementing updates. Creation and deletion plans are pretty straightforward; we just take a single graph, topologically sort it, and perform the operations in the right order. For creation, this is in dependency order (things that are depended upon must be created before dependents); for deletion, this is in reverse-dependency order (things that depend on others must be deleted before dependencies). These are just special cases of the more general idea of performing DAG operations in dependency order. Updates must work in terms of this more general notion. For example: * It is an error to delete a resource while another refers to it; thus, resources are deleted after deleting dependents, or after updating dependent properties that reference the resource to new values. * It is an error to depend on a create a resource before it is created; thus, resources must be created before dependents are created, and/or before updates to existing resource properties that would cause them to refer to the new resource. Of course, all of this is tangled up in a graph of dependencies. As a result, we must create a DAG of the dependencies between creates, updates, and deletes, and then topologically sort this DAG, in order to determine the proper order of update operations. To do this, we slightly generalize the existing graph infrastructure, while also specializing two kinds of graphs; the existing one becomes a heapstate.ObjectGraph, while this new one is resource.planGraph (internal).
2017-02-23 23:56:23 +01:00
ov := v.(*heapstate.ObjectVertex)
if IsResourceVertex(ov) {
Redo object monikers This change overhauls the way we do object monikers. The old mechanism, generating monikers using graph paths, was far too brittle and prone to collisions. The new approach mixes some amount of "automatic scoping" plus some "explicit naming." Although there is some explicitness, this is arguably a good thing, as the monikers will be relatable back to the source more readily by developers inspecting the graph and resource state. Each moniker has four parts: <Namespace>::<AllocModule>::<Type>::<Name> wherein each element is the following: <Namespace> The namespace being deployed into <AllocModule> The module in which the object was allocated <Type> The type of the resource <Name> The assigned name of the resource The <Namespace> is essentially the deployment target -- so "prod", "stage", etc -- although it is more general purpose to allow for future namespacing within a target (e.g., "prod/customer1", etc); for now this is rudimentary, however, see marapongo/mu#94. The <AllocModule> is the token for the code that contained the 'new' that led to this object being created. In the future, we may wish to extend this to also track the module under evaluation. (This is a nice aspect of monikers; they can become arbitrarily complex, so long as they are precise, and not prone to false positives/negatives.) The <Name> warrants more discussion. The resource provider is consulted via a new gRPC method, Name, that fetches the name. How the provider does this is entirely up to it. For some resource types, the resource may have properties that developers must set (e.g., `new Bucket("foo")`); for other providers, perhaps the resource intrinsically has a property that explicitly and uniquely qualifies the object (e.g., AWS SecurityGroups, via `new SecurityGroup({groupName: "my-sg"}`); and finally, it's conceivable that a provider might auto-generate the name (e.g., such as an AWS Lambda whose name could simply be a hash of the source code contents). This should overall produce better results with respect to moniker collisions, ability to match resources, and the usability of the system.
2017-02-24 23:50:02 +01:00
resobjs = append(resobjs, ov.Obj())
Begin resource modeling and planning This change introduces a new package, pkg/resource, that will form the foundation for actually performing deployment plans and applications. It contains the following key abstractions: * resource.Provider is a wrapper around the CRUD operations exposed by underlying resource plugins. It will eventually defer to resource.Plugin, which itself defers -- over an RPC interface -- to the actual plugin, one per package exposing resources. The provider will also understand how to load, cache, and overall manage the lifetime of each plugin. * resource.Resource is the actual resource object. This is created from the overall evaluation object graph, but is simplified. It contains only serializable properties, for example. Inter-resource references are translated into serializable monikers as part of creating the resource. * resource.Moniker is a serializable string that uniquely identifies a resource in the Mu system. This is in contrast to resource IDs, which are generated by resource providers and generally opaque to the Mu system. See marapongo/mu#69 for more information about monikers and some of their challenges (namely, designing a stable algorithm). * resource.Snapshot is a "snapshot" taken from a graph of resources. This is a transitive closure of state representing one possible configuration of a given environment. This is what plans are created from. Eventually, two snapshots will be diffable, in order to perform incremental updates. One way of thinking about this is that a snapshot of the old world's state is advanced, one step at a time, until it reaches a desired snapshot of the new world's state. * resource.Plan is a plan for carrying out desired CRUD operations on a target environment. Each plan consists of zero-to-many Steps, each of which has a CRUD operation type, a resource target, and a next step. This is an enumerator because it is possible the plan will evolve -- and introduce new steps -- as it is carried out (hence, the Next() method). At the moment, this is linearized; eventually, we want to make this more "graph-like" so that we can exploit available parallelism within the dependencies. There are tons of TODOs remaining. However, the `mu plan` command is functioning with these new changes -- including colorization FTW -- so I'm landing it now. This is part of marapongo/mu#38 and marapongo/mu#41.
2017-02-17 21:31:48 +01:00
}
}
Redo object monikers This change overhauls the way we do object monikers. The old mechanism, generating monikers using graph paths, was far too brittle and prone to collisions. The new approach mixes some amount of "automatic scoping" plus some "explicit naming." Although there is some explicitness, this is arguably a good thing, as the monikers will be relatable back to the source more readily by developers inspecting the graph and resource state. Each moniker has four parts: <Namespace>::<AllocModule>::<Type>::<Name> wherein each element is the following: <Namespace> The namespace being deployed into <AllocModule> The module in which the object was allocated <Type> The type of the resource <Name> The assigned name of the resource The <Namespace> is essentially the deployment target -- so "prod", "stage", etc -- although it is more general purpose to allow for future namespacing within a target (e.g., "prod/customer1", etc); for now this is rudimentary, however, see marapongo/mu#94. The <AllocModule> is the token for the code that contained the 'new' that led to this object being created. In the future, we may wish to extend this to also track the module under evaluation. (This is a nice aspect of monikers; they can become arbitrarily complex, so long as they are precise, and not prone to false positives/negatives.) The <Name> warrants more discussion. The resource provider is consulted via a new gRPC method, Name, that fetches the name. How the provider does this is entirely up to it. For some resource types, the resource may have properties that developers must set (e.g., `new Bucket("foo")`); for other providers, perhaps the resource intrinsically has a property that explicitly and uniquely qualifies the object (e.g., AWS SecurityGroups, via `new SecurityGroup({groupName: "my-sg"}`); and finally, it's conceivable that a provider might auto-generate the name (e.g., such as an AWS Lambda whose name could simply be a hash of the source code contents). This should overall produce better results with respect to moniker collisions, ability to match resources, and the usability of the system.
2017-02-24 23:50:02 +01:00
return resobjs, nil
Begin resource modeling and planning This change introduces a new package, pkg/resource, that will form the foundation for actually performing deployment plans and applications. It contains the following key abstractions: * resource.Provider is a wrapper around the CRUD operations exposed by underlying resource plugins. It will eventually defer to resource.Plugin, which itself defers -- over an RPC interface -- to the actual plugin, one per package exposing resources. The provider will also understand how to load, cache, and overall manage the lifetime of each plugin. * resource.Resource is the actual resource object. This is created from the overall evaluation object graph, but is simplified. It contains only serializable properties, for example. Inter-resource references are translated into serializable monikers as part of creating the resource. * resource.Moniker is a serializable string that uniquely identifies a resource in the Mu system. This is in contrast to resource IDs, which are generated by resource providers and generally opaque to the Mu system. See marapongo/mu#69 for more information about monikers and some of their challenges (namely, designing a stable algorithm). * resource.Snapshot is a "snapshot" taken from a graph of resources. This is a transitive closure of state representing one possible configuration of a given environment. This is what plans are created from. Eventually, two snapshots will be diffable, in order to perform incremental updates. One way of thinking about this is that a snapshot of the old world's state is advanced, one step at a time, until it reaches a desired snapshot of the new world's state. * resource.Plan is a plan for carrying out desired CRUD operations on a target environment. Each plan consists of zero-to-many Steps, each of which has a CRUD operation type, a resource target, and a next step. This is an enumerator because it is possible the plan will evolve -- and introduce new steps -- as it is carried out (hence, the Next() method). At the moment, this is linearized; eventually, we want to make this more "graph-like" so that we can exploit available parallelism within the dependencies. There are tons of TODOs remaining. However, the `mu plan` command is functioning with these new changes -- including colorization FTW -- so I'm landing it now. This is part of marapongo/mu#38 and marapongo/mu#41.
2017-02-17 21:31:48 +01:00
}