Documentation
¶
Index ¶
Constants ¶
const ( // MachineScope is a special scope name that is used // for machine placement directives (e.g. --to 0). MachineScope = "#" )
Variables ¶
var ContainerTypes = []ContainerType{ LXD, KVM, }
ContainerTypes is used to validate add-machine arguments.
var ErrPlacementScopeMissing = fmt.Errorf("placement scope missing")
Functions ¶
This section is empty.
Types ¶
type ContainerType ¶
type ContainerType string
ContainerType defines different container technologies known to juju.
const ( NONE ContainerType = "none" LXD ContainerType = "lxd" KVM ContainerType = "kvm" )
Known container types.
func ParseContainerType ¶
func ParseContainerType(ctype string) (ContainerType, error)
ParseContainerType converts the specified string into a supported ContainerType instance or returns an error if the container type is invalid.
func ParseContainerTypeOrNone ¶
func ParseContainerTypeOrNone(ctype string) (ContainerType, error)
ParseContainerTypeOrNone converts the specified string into a supported ContainerType instance or returns an error if the container type is invalid. For this version of the function, 'none' is a valid value.
type Distributor ¶
type Distributor interface {
// DistributeInstance takes a set of clean, empty
// instances, and a distribution group, and returns
// the subset of candidates which the policy will
// allow entry into the distribution group.
//
// The AssignClean and AssignCleanEmpty unit
// assignment policies will attempt to assign a
// unit to each of the resulting instances until
// one is successful. If no instances can be assigned
// to (e.g. because of concurrent deployments), then
// a new machine will be allocated.
DistributeInstances(candidates, distributionGroup []Id) ([]Id, error)
}
Distributor is an interface that may be used to distribute application units across instances for high availability.
type HardwareCharacteristics ¶
type HardwareCharacteristics struct {
// Arch is the architecture of the processor.
Arch *string `json:"arch,omitempty" yaml:"arch,omitempty"`
// Mem is the size of RAM in megabytes.
Mem *uint64 `json:"mem,omitempty" yaml:"mem,omitempty"`
// RootDisk is the size of the disk in megabytes.
RootDisk *uint64 `json:"root-disk,omitempty" yaml:"rootdisk,omitempty"`
// CpuCores is the number of logical cores the processor has.
CpuCores *uint64 `json:"cpu-cores,omitempty" yaml:"cpucores,omitempty"`
// CpuPower is a relative representation of the speed of the processor.
CpuPower *uint64 `json:"cpu-power,omitempty" yaml:"cpupower,omitempty"`
// Tags is a list of strings that identify the machine.
Tags *[]string `json:"tags,omitempty" yaml:"tags,omitempty"`
// AvailabilityZone defines the zone in which the machine resides.
AvailabilityZone *string `json:"availability-zone,omitempty" yaml:"availabilityzone,omitempty"`
}
HardwareCharacteristics represents the characteristics of the instance (if known). Attributes that are nil are unknown or not supported.
func MustParseHardware ¶
func MustParseHardware(args ...string) HardwareCharacteristics
MustParseHardware constructs a HardwareCharacteristics from the supplied arguments, as Parse, but panics on failure.
func ParseHardware ¶
func ParseHardware(args ...string) (HardwareCharacteristics, error)
ParseHardware constructs a HardwareCharacteristics from the supplied arguments, each of which must contain only spaces and name=value pairs. If any name is specified more than once, an error is returned.
func (HardwareCharacteristics) String ¶
func (hc HardwareCharacteristics) String() string
type Id ¶
type Id string
An instance Id is a provider-specific identifier associated with an instance (physical or virtual machine allocated in the provider).
const UnknownId Id = ""
UnknownId can be used to explicitly specify the instance ID does not matter.
type Instance ¶
type Instance interface {
// Id returns a provider-generated identifier for the Instance.
Id() Id
// Status returns the provider-specific status for the instance.
Status() InstanceStatus
// Addresses returns a list of hostnames or ip addresses
// associated with the instance.
Addresses() ([]network.Address, error)
// OpenPorts opens the given port ranges on the instance, which
// should have been started with the given machine id.
OpenPorts(machineId string, rules []network.IngressRule) error
// ClosePorts closes the given port ranges on the instance, which
// should have been started with the given machine id.
ClosePorts(machineId string, rules []network.IngressRule) error
// IngressRules returns the set of ingress rules for the instance,
// which should have been applied to the given machine id. The
// rules are returned as sorted by network.SortIngressRules().
// It is expected that there be only one ingress rule result for a given
// port range - the rule's SourceCIDRs will contain all applicable source
// address rules for that port range.
IngressRules(machineId string) ([]network.IngressRule, error)
}
Instance represents the the realization of a machine in state.
type InstanceStatus ¶
InstanceStatus represents the status for a provider instance.
type Namespace ¶
type Namespace interface {
// Prefix returns the common part of the hostnames. i.e. 'juju-xxxxxx-'
Prefix() string
// Hostname returns a name suitable to be used for a machine hostname.
// This function returns an error if the machine tags is invalid.
Hostname(machineID string) (string, error)
// MachineTag does the reverse of the Hostname method, and extracts the
// Tag from the hostname.
MachineTag(hostname string) (names.MachineTag, error)
// Value returns the input prefixed with the namespace prefix.
Value(string) string
}
Namespace provides a way to generate machine hostanmes with a given prefix.
func NewNamespace ¶
NewNamespace returns a Namespace identified by the last six hex digits of the model UUID. NewNamespace returns an error if the model tag is invalid.
type Placement ¶
type Placement struct {
// Scope is the scope of the placement directive. Scope may
// be a container type (lxd, kvm), instance.MachineScope, or
// an environment name.
//
// If Scope is empty, then it must be inferred from the context.
Scope string `json:"scope"`
// Directive is a scope-specific placement directive.
//
// For MachineScope or a container scope, this may be empty or
// the ID of an existing machine.
Directive string `json:"directive"`
}
Placement defines a placement directive, which has a scope and a value that is scope-specific.
func MustParsePlacement ¶
MustParsePlacement attempts to parse the specified string and create a corresponding Placement structure, panicking if an error occurs.
func ParsePlacement ¶
ParsePlacement attempts to parse the specified string and create a corresponding Placement structure.
If the placement directive is non-empty and missing a scope, ErrPlacementScopeMissing will be returned as well as a Placement with an empty Scope field.