Documentation
¶
Overview ¶
Package types describes few of the essential types used throughout the application.
Package types describes few of the essential types used throughout the application.
Index ¶
- Constants
- Variables
- type App
- type AppStats
- type AppVersion
- type BytesSize
- type CacheAlgorithm
- type CacheStats
- type CacheZone
- type FilePath
- type Location
- type Logger
- type NilNextHandler
- type NotNilNextHandler
- type ObjectID
- type ObjectIDHash
- type ObjectIndex
- type ObjectIndexHash
- type ObjectMetadata
- type RequestHandler
- type RequestHandlerFunc
- type Scheduler
- type Storage
- type Upstream
- type UpstreamAddress
- type UpstreamBalancingAlgorithm
Constants ¶
const ObjectIDHashSize = sha1.Size
ObjectIDHashSize is the size of the byte array that contains the object hash.
const ObjectIndexHashSize = ObjectIDHashSize + 4 // sizeof(uint32)
ObjectIndexHashSize is the size of the byte array that contains the object hash.
Variables ¶
var (
ErrAlreadyInCache = errors.New("Object already in cache")
)
Exported errors
Functions ¶
This section is empty.
Types ¶
type App ¶
type App interface {
// Stats returns applicaiton wide stats
Stats() AppStats
// Started returns the time at which the app was started
Started() time.Time
// Version returns a string representation of the version of the app
Version() AppVersion
// GetLocationFor returns the Location that mathes the provided host and path
GetLocationFor(host, path string) *Location
// GetUpstream gets an upstream by it's id, nil is returned if no such is defined
GetUpstream(id string) Upstream
}
App is an interface for an application to implement
type AppStats ¶
type AppStats struct {
Requests, Responded, NotConfigured uint64
}
AppStats are stats for the whole application
type AppVersion ¶
type AppVersion struct {
Dirty bool
Version string
GitHash string
GitTag string
BuildTime time.Time
}
AppVersion is struct representing an App version
func (AppVersion) String ¶
func (a AppVersion) String() string
type BytesSize ¶
type BytesSize uint64
BytesSize represents size written in string format. Examples: "1m", "20g" etc. Its main purpose is to be stored and loaded from json.
func BytesSizeFromString ¶
BytesSizeFromString parses bytes size such as "1m", "15g" to BytesSize struct.
func (*BytesSize) MarshalJSON ¶
MarshalJSON is needed for automatic marshalling of BytesSize fields in the JSON configuration.
func (*BytesSize) UnmarshalJSON ¶
UnmarshalJSON is needed for automatic unmarshalling of BytesSize fields in the JSON configuration.
type CacheAlgorithm ¶
type CacheAlgorithm interface {
// Lookup returns wheather this object is in the cache or not
Lookup(*ObjectIndex) bool
// ShouldKeep is called to signal that this ObjectIndex has been stored
ShouldKeep(*ObjectIndex) bool
// AddObject adds this ObjectIndex to the cache. Returns an error when
// the object is in the cache already.
AddObject(*ObjectIndex) error
// PromoteObject is called every time this part of a file has been used
// to satisfy a client request
PromoteObject(*ObjectIndex)
// ConsumedSize returns the full size of all files currently in the cache
ConsumedSize() BytesSize
// Stats returns statistics for this cache algorithm
Stats() CacheStats
// Remove all of the provided object indexes from the cache.
Remove(...*ObjectIndex)
// ChangeConfig changes the changable parts of the a CacheAlgorithm:
// the timeout and count for removing objects in bulk, the logger it uses
// and the count of objects it contains. Automatically resizing the algorithm
// if it's required
ChangeConfig(bulkTimeout, bulkCount, objectCount uint64, l Logger)
}
CacheAlgorithm interface defines how a cache should behave
type CacheStats ¶
type CacheStats interface {
// CacheHitPrc returns a string such as '53%' which represents the cache hit
// ratio of this cache. Basically this number is (Hits()/Requests()) * 100.
CacheHitPrc() string
// ID is a way of identifing this particular cache zone
ID() string
// Hits returns the number of cache hits this cache has generated
Hits() uint64
// Requests returns the number of lookups in the cache
Requests() uint64
// Objects returns the number of cache object at the moment. These are the actual
// objects on the disk. Not maximum possible objects
Objects() uint64
// Size returns the consumed space in bytes for this cache
Size() BytesSize
}
CacheStats is the common interface that is used so every cache can generate a Stats object which is to be used in the server status page.
type CacheZone ¶
type CacheZone struct {
ID string
PartSize BytesSize
Algorithm CacheAlgorithm
Scheduler Scheduler
Storage Storage
}
CacheZone is the combination of a Storage for storing object parts and an `CacheAlgorithm` which determines what should be stored.
type FilePath ¶
type FilePath string
FilePath represents the string path to a file on the local filesystem. It is mainly used for flag arguments because it implements the flag.Value interface and resolves the absolute path to the file when the Set() method is used. Thus, if the current working directory is changed during the execution of the program, the path still remains valid.
type Location ¶
type Location struct {
Name string
Handler RequestHandler
CacheKey string
CacheDefaultDuration time.Duration
CacheKeyIncludesQuery bool
Cache *CacheZone //!TODO: move to the cache handler settings (plus all Cache* settings)
Upstream Upstream
Logger Logger
}
Location links a config location to its cache algorithm and a storage object.
func (*Location) NewObjectIDForURL ¶
NewObjectIDForURL returns new ObjectID from the provided URL
type Logger ¶
type Logger interface {
// Log emits this message to the log stream.
Log(v ...interface{})
// Logf emits this message to the log stream. Supports fmt.Printf formating.
Logf(format string, args ...interface{})
// Debug emits this message to the debug stream.
Debug(v ...interface{})
// Debuggerebugf emits this message to the debug stream. Supports fmt.Printf formatting.
Debugf(format string, args ...interface{})
// Error emits this message to the error stream.
Error(v ...interface{})
// Errorf emits this message to the error stream. Supports fmt.Printf formatting.
Errorf(format string, args ...interface{})
// Fatal will print the message to the error stream and halt the program.
Fatal(v ...interface{})
// Fatalf is similar to Fatal but supports formatting. Support fmt.Printf formating.
Fatalf(format string, v ...interface{})
}
Logger is the common interface that all nedomi loggers should implement.
type NilNextHandler ¶
type NilNextHandler string
NilNextHandler is an error type to be returned when a next handler is required but it's nil
func (NilNextHandler) Error ¶
func (n NilNextHandler) Error() string
type NotNilNextHandler ¶
type NotNilNextHandler string
NotNilNextHandler is an error type to be returned when a next handler is not nil but it should be
func (NotNilNextHandler) Error ¶
func (n NotNilNextHandler) Error() string
type ObjectID ¶
type ObjectID struct {
// contains filtered or unexported fields
}
ObjectID represents a cached file.
func NewObjectID ¶
NewObjectID creates and returns a new ObjectID.
func (*ObjectID) Hash ¶
func (oid *ObjectID) Hash() ObjectIDHash
Hash returns the pre-calculated sha1 hash of the object id.
func (*ObjectID) MarshalJSON ¶
MarshalJSON is used to help the JSON library marshal the unexported vars.
func (*ObjectID) UnmarshalJSON ¶
UnmarshalJSON is used to help the JSON library unmarshal the unexported vars.
type ObjectIDHash ¶
type ObjectIDHash [ObjectIDHashSize]byte
ObjectIDHash is the fixed-width byte array that represents an ObjectID hash.
type ObjectIndex ¶
ObjectIndex represents a particular index in a file
func (*ObjectIndex) Hash ¶
func (oi *ObjectIndex) Hash() ObjectIndexHash
Hash returns the pre-calculated sha1 of the ObjectID with concatinated the 4 bytes identifying the part
func (*ObjectIndex) String ¶
func (oi *ObjectIndex) String() string
type ObjectIndexHash ¶
type ObjectIndexHash [ObjectIndexHashSize]byte
ObjectIndexHash is the fixed-width byte array that represents an ObjectID hash.
type ObjectMetadata ¶
type ObjectMetadata struct {
// The ObjectID for this file. It is used for identifying the object throughout
// the cache layers in nedomi.
ID *ObjectID
// The time of the first request/response for this object as unix timestamp.
ResponseTimestamp int64
// Status code of the first proxied response for this object.
Code int
// The object size in bytes. Normally this should correspond to the
// upstream's Content-Length header.
Size uint64
// HTTP headers which were received from the upstream and which we should
// pass down for this object for any subsequent request.
Headers http.Header
// The time at wich this object can be considered stale. After this time
// the object must be revalidated or discarded. This value is a unix timestamp.
ExpiresAt int64
}
ObjectMetadata represents all the needed metadata of a cacheable object.
type RequestHandler ¶
type RequestHandler interface {
// RequestHandle is function similar to the http.ServeHTTP. It differs only
// in that it has context as an extra argument.
RequestHandle(context.Context, http.ResponseWriter, *http.Request)
}
RequestHandler interface defines the RequestHandle funciton. All nedomi handle modules must implement this interface.
type RequestHandlerFunc ¶
The RequestHandlerFunc type is an adapter to allow the use of ordinary functions as request handlers. If f is a function with the appropriate signature, HandlerFunc(f) is a Handler object that calls f.
func (RequestHandlerFunc) RequestHandle ¶
func (rhf RequestHandlerFunc) RequestHandle(ctx context.Context, w http.ResponseWriter, req *http.Request)
RequestHandle with rhf(ctx, w, req)
type Scheduler ¶
type Scheduler interface {
// AddEvent schedules the passed callback to be executed at the supplied time.
AddEvent(key ObjectIDHash, callback func(), in time.Duration)
// Contains checks whether an event with the supplied key is scheduled.
Contains(key ObjectIDHash) bool
}
Scheduler efficiently manages and executes callbacks at specified times.
type Storage ¶
type Storage interface {
// Returns the maximum part size for the storage.
PartSize() uint64
// Returns the metadata for this object, it it is present. If the requested
// metadata is not on the storage, it returns os.ErrNotExist.
GetMetadata(id *ObjectID) (*ObjectMetadata, error)
// Returns an io.ReadCloser instance that will read the specified part of
// the object, if it is present. If the requested part is not on the
// storage, it will return os.ErrNotExist.
GetPart(id *ObjectIndex) (io.ReadCloser, error)
// GetAvailableParts returns types.ObjectIndexMap including all the available
// parts of for the object specified by the provided objectMetadata
GetAvailableParts(id *ObjectID) ([]*ObjectIndex, error)
// Saves the supplied metadata to the storage.
SaveMetadata(m *ObjectMetadata) error
// Saves the contents of the supplied object part to the storage.
SavePart(index *ObjectIndex, data io.Reader) error
// Discard an object and its metadata from the storage.
Discard(id *ObjectID) error
// Discard the specified part of an Object from the storage.
DiscardPart(index *ObjectIndex) error
// Iterate iterates over the storage objects and passes them and information
// about their parts to the supplied callback function. It is used for
// restoring the state after the service has been restarted. When the
// callback returns false, the iteration stops.
Iterate(callback func(*ObjectMetadata, ...*ObjectIndex) bool) error
// ChangeConfig change configs of the storage and start using them
ChangeConfig(Logger)
}
Storage represents a single unit of storage.
type Upstream ¶
type Upstream interface {
http.RoundTripper
CancelRequest(*http.Request)
GetAddress(string) (*UpstreamAddress, error)
}
Upstream represents an object that is used by the proxy handler for making requests to the configured upstream server or servers.
type UpstreamAddress ¶
type UpstreamAddress struct {
url.URL
Hostname string
Port string
OriginalURL *url.URL
Weight uint32
}
UpstreamAddress represents a resolved upstream address
func (*UpstreamAddress) String ¶ added in v0.1.2
func (ua *UpstreamAddress) String() string
type UpstreamBalancingAlgorithm ¶
type UpstreamBalancingAlgorithm interface {
// Set is used to specify all the possible upstream addresses the algorithm
// has to choose from.
Set([]*UpstreamAddress)
// Get returns a specific address, according to the supplied path.
Get(string) (*UpstreamAddress, error)
}
UpstreamBalancingAlgorithm encapulates thread-safe methods that are used for balancing user requests between a set of upstream addresses. That is done according to the specific balancing algorithm implementation.
Source Files