driver

package
v1.2.0 Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Dec 10, 2019 License: Apache-2.0 Imports: 26 Imported by: 0

Documentation

Index

Constants

This section is empty.

Variables

View Source 
var (
	// UnknownTransactionCommitResult is an error label for unknown transaction commit results.
	UnknownTransactionCommitResult = "UnknownTransactionCommitResult"
	// TransientTransactionError is an error label for transient errors with transactions.
	TransientTransactionError = "TransientTransactionError"
	// NetworkError is an error label for network errors.
	NetworkError = "NetworkError"
	// ErrCursorNotFound is the cursor not found error for legacy find operations.
	ErrCursorNotFound = errors.New("cursor not found")
	// ErrUnacknowledgedWrite is returned from functions that have an unacknowledged
	// write concern.
	ErrUnacknowledgedWrite = errors.New("unacknowledged write")
	// ErrUnsupportedStorageEngine is returned when a retryable write is attempted against a server
	// that uses a storage engine that does not support retryable writes
	ErrUnsupportedStorageEngine = errors.New("this MongoDB deployment does not support retryable writes. Please add retryWrites=false to your connection string")
)
View Source 
var (
	// ErrNoDocCommandResponse occurs when the server indicated a response existed, but none was found.
	ErrNoDocCommandResponse = errors.New("command returned no documents")
	// ErrMultiDocCommandResponse occurs when the server sent multiple documents in response to a command.
	ErrMultiDocCommandResponse = errors.New("command returned multiple documents")
	// ErrReplyDocumentMismatch occurs when the number of documents returned in an OP_QUERY does not match the numberReturned field.
	ErrReplyDocumentMismatch = errors.New("number of documents returned does not match numberReturned field")
	// ErrNonPrimaryReadPref is returned when a read is attempted in a transaction with a non-primary read preference.
	ErrNonPrimaryReadPref = errors.New("read preference in a transaction must be primary")
)
View Source 
var ErrDocumentTooLarge = errors.New("an inserted document is too large")

ErrDocumentTooLarge occurs when a document that is larger than the maximum size accepted by a server is passed to an insert command.

View Source 
var (

	// ErrFilterType is returned when the filter for a legacy list collections operation is of the wrong type.
	ErrFilterType = errors.New("filter for list collections operation must be a string")
)

Functions

func CompressPlayoad added in v1.2.0 

func CompressPlayoad(in []byte, opts CompressionOpts) ([]byte, error)

CompressPlayoad takes a byte slice and compresses it according to the options passed

func DecompressPayload added in v1.2.0 

func DecompressPayload(in []byte, opts CompressionOpts) ([]byte, error)

DecompressPayload takes a byte slice that has been compressed and undoes it according to the options passed

Types

type BatchCursor added in v0.3.0 

type BatchCursor struct {
	// contains filtered or unexported fields
}

BatchCursor is a batch implementation of a cursor. It returns documents in entire batches instead of one at a time. An individual document cursor can be built on top of this batch cursor.

func NewBatchCursor added in v0.3.0 

func NewBatchCursor(cr CursorResponse, clientSession *session.Client, clock *session.ClusterClock, opts CursorOptions) (*BatchCursor, error)

NewBatchCursor creates a new BatchCursor from the provided parameters.

func NewEmptyBatchCursor added in v0.3.0 

func NewEmptyBatchCursor() *BatchCursor

NewEmptyBatchCursor returns a batch cursor that is empty.

func (*BatchCursor) Batch added in v0.3.0 

func (bc *BatchCursor) Batch() *bsoncore.DocumentSequence

Batch will return a DocumentSequence for the current batch of documents. The returned DocumentSequence is only valid until the next call to Next or Close.

func (*BatchCursor) Close added in v0.3.0 

func (bc *BatchCursor) Close(ctx context.Context) error

Close closes this batch cursor.

func (*BatchCursor) Err added in v0.3.0 

func (bc *BatchCursor) Err() error

Err returns the latest error encountered.

func (*BatchCursor) ID added in v0.3.0 

func (bc *BatchCursor) ID() int64

ID returns the cursor ID for this batch cursor.

func (*BatchCursor) KillCursor added in v1.1.0 

func (bc *BatchCursor) KillCursor(ctx context.Context) error

KillCursor kills cursor on server without closing batch cursor

func (*BatchCursor) Next added in v0.3.0 

func (bc *BatchCursor) Next(ctx context.Context) bool

Next indicates if there is another batch available. Returning false does not necessarily indicate that the cursor is closed. This method will return false when an empty batch is returned.

If Next returns true, there is a valid batch of documents available. If Next returns false, there is not a valid batch of documents available.

func (*BatchCursor) PostBatchResumeToken added in v1.1.0 

func (bc *BatchCursor) PostBatchResumeToken() bsoncore.Document

PostBatchResumeToken returns the latest seen post batch resume token.

func (*BatchCursor) Server added in v1.0.0 

func (bc *BatchCursor) Server() Server

Server returns the server for this cursor.

type Batches added in v1.1.0 

type Batches struct {
	Identifier string
	Documents  []bsoncore.Document
	Current    []bsoncore.Document
	Ordered    *bool
}

Batches contains the necessary information to batch split an operation. This is only used for write oeprations.

func (*Batches) AdvanceBatch added in v1.1.0 

func (b *Batches) AdvanceBatch(maxCount, targetBatchSize, maxDocSize int) error

AdvanceBatch splits the next batch using maxCount and targetBatchSize. This method will do nothing if the current batch has not been cleared. We do this so that when this is called during execute we can call it without first needing to check if we already have a batch, which makes the code simpler and makes retrying easier. The maxDocSize parameter is used to check that any one document is not too large. If the first document is bigger than targetBatchSize but smaller than maxDocSize, a batch of size 1 containing that document will be created.

func (*Batches) ClearBatch added in v1.1.0 

func (b *Batches) ClearBatch()

ClearBatch clears the Current batch. This must be called before AdvanceBatch will advance to the next batch.

func (*Batches) Valid added in v1.1.0 

func (b *Batches) Valid() bool

Valid returns true if Batches contains both an identifier and the length of Documents is greater than zero.

type CollectionInfoFn added in v1.2.0 

type CollectionInfoFn func(ctx context.Context, db string, filter bsoncore.Document) (bsoncore.Document, error)

CollectionInfoFn is a callback used to retrieve collection information.

type CompressionOpts added in v1.2.0 

type CompressionOpts struct {
	Compressor       wiremessage.CompressorID
	ZlibLevel        int
	ZstdLevel        int
	UncompressedSize int32
}

CompressionOpts holds settings for how to compress a payload

type Compressor added in v1.1.0 

type Compressor interface {
	CompressWireMessage(src, dst []byte) ([]byte, error)
}

Compressor is an interface used to compress wire messages. If a Connection supports compression it should implement this interface as well. The CompressWireMessage method will be called during the execution of an operation if the wire message is allowed to be compressed.

type Connection added in v1.1.0 

type Connection interface {
	WriteWireMessage(context.Context, []byte) error
	ReadWireMessage(ctx context.Context, dst []byte) ([]byte, error)
	Description() description.Server
	Close() error
	ID() string
	Address() address.Address
}

Connection represents a connection to a MongoDB server.

type Connector added in v1.2.0 

type Connector interface {
	Connect() error
}

Connector represents a type that can connect to a server.

type Crypt added in v1.2.0 

type Crypt struct {
	BypassAutoEncryption bool
	// contains filtered or unexported fields
}

Crypt consumes the libmongocrypt.MongoCrypt type to iterate the mongocrypt state machine and perform encryption and decryption.

func NewCrypt added in v1.2.0 

func NewCrypt(opts *CryptOptions) (*Crypt, error)

NewCrypt creates a new Crypt instance configured with the given AutoEncryptionOptions.

func (*Crypt) Close added in v1.2.0 

func (c *Crypt) Close()

Close cleans up any resources associated with the Crypt instance.

func (*Crypt) CreateDataKey added in v1.2.0 

func (c *Crypt) CreateDataKey(ctx context.Context, kmsProvider string, opts *options.DataKeyOptions) (bsoncore.Document, error)

CreateDataKey creates a data key using the given KMS provider and options.

func (*Crypt) Decrypt added in v1.2.0 

func (c *Crypt) Decrypt(ctx context.Context, cmdResponse bsoncore.Document) (bsoncore.Document, error)

Decrypt decrypts the given command response.

func (*Crypt) DecryptExplicit added in v1.2.0 

func (c *Crypt) DecryptExplicit(ctx context.Context, subtype byte, data []byte) (bsoncore.Value, error)

DecryptExplicit decrypts the given encrypted value.

func (*Crypt) Encrypt added in v1.2.0 

func (c *Crypt) Encrypt(ctx context.Context, db string, cmd bsoncore.Document) (bsoncore.Document, error)

Encrypt encrypts the given command.

func (*Crypt) EncryptExplicit added in v1.2.0 

func (c *Crypt) EncryptExplicit(ctx context.Context, val bsoncore.Value, opts *options.ExplicitEncryptionOptions) (byte, []byte, error)

EncryptExplicit encrypts the given value with the given options.

type CryptOptions added in v1.2.0 

type CryptOptions struct {
	CollInfoFn           CollectionInfoFn
	KeyFn                KeyRetrieverFn
	MarkFn               MarkCommandFn
	KmsProviders         map[string]map[string]interface{}
	SchemaMap            map[string]bsoncore.Document
	BypassAutoEncryption bool
}

CryptOptions specifies options to configure a Crypt instance.

type CursorOptions added in v1.1.0 

type CursorOptions struct {
	BatchSize      int32
	MaxTimeMS      int64
	Limit          int32
	CommandMonitor *event.CommandMonitor
	Crypt          *Crypt
}

CursorOptions are extra options that are required to construct a BatchCursor.

type CursorResponse added in v1.1.0 

type CursorResponse struct {
	Server     Server
	Desc       description.Server
	FirstBatch *bsoncore.DocumentSequence
	Database   string
	Collection string
	ID         int64
	// contains filtered or unexported fields
}

CursorResponse represents the response from a command the results in a cursor. A BatchCursor can be constructed from a CursorResponse.

func NewCursorResponse added in v1.1.0 

func NewCursorResponse(response bsoncore.Document, server Server, desc description.Server) (CursorResponse, error)

NewCursorResponse constructs a cursor response from the given response and server. This method can be used within the ProcessResponse method for an operation.

type Deployment added in v1.1.0 

type Deployment interface {
	SelectServer(context.Context, description.ServerSelector) (Server, error)
	SupportsRetryWrites() bool
	Kind() description.TopologyKind
}

Deployment is implemented by types that can select a server from a deployment.

type Disconnector added in v1.2.0 

type Disconnector interface {
	Disconnect(context.Context) error
}

Disconnector represents a type that can disconnect from a server.

type Error added in v1.1.0 

type Error struct {
	Code    int32
	Message string
	Labels  []string
	Name    string
}

Error is a command execution error from the database.

func (Error) Error added in v1.1.0 

func (e Error) Error() string

Error implements the error interface.

func (Error) HasErrorLabel added in v1.1.0 

func (e Error) HasErrorLabel(label string) bool

HasErrorLabel returns true if the error contains the specified label.

func (Error) NamespaceNotFound added in v1.1.0 

func (e Error) NamespaceNotFound() bool

NamespaceNotFound returns true if this errors is a NamespaceNotFound error.

func (Error) NetworkError added in v1.1.0 

func (e Error) NetworkError() bool

NetworkError returns true if the error is a network error.

func (Error) NodeIsRecovering added in v1.1.0 

func (e Error) NodeIsRecovering() bool

NodeIsRecovering returns true if this error is a node is recovering error.

func (Error) NodeIsShuttingDown added in v1.1.0 

func (e Error) NodeIsShuttingDown() bool

NodeIsShuttingDown returns true if this error is a node is shutting down error.

func (Error) NotMaster added in v1.1.0 

func (e Error) NotMaster() bool

NotMaster returns true if this error is a not master error.

func (Error) Retryable added in v1.1.0 

func (e Error) Retryable() bool

Retryable returns true if the error is retryable

func (Error) UnsupportedStorageEngine added in v1.1.0 

func (e Error) UnsupportedStorageEngine() bool

UnsupportedStorageEngine returns whether e came as a result of an unsupported storage engine

type ErrorProcessor added in v1.1.0 

type ErrorProcessor interface {
	ProcessError(error)
}

ErrorProcessor implementations can handle processing errors, which may modify their internal state. If this type is implemented by a Server, then Operation.Execute will call it's ProcessError method after it decodes a wire message.

type Expirable added in v1.1.0 

type Expirable interface {
	Expire() error
	Alive() bool
}

Expirable represents an expirable object.

type Handshaker added in v1.1.0 

type Handshaker interface {
	GetDescription(context.Context, address.Address, Connection) (description.Server, error)
	FinishHandshake(context.Context, Connection) error
}

Handshaker is the interface implemented by types that can perform a MongoDB handshake over a provided driver.Connection. This is used during connection initialization. Implementations must be goroutine safe.

type InvalidOperationError added in v1.1.0 

type InvalidOperationError struct{ MissingField string }

InvalidOperationError is returned from Validate and indicates that a required field is missing from an instance of Operation.

func (InvalidOperationError) Error added in v1.1.0 

func (err InvalidOperationError) Error() string

type KeyRetrieverFn added in v1.2.0 

type KeyRetrieverFn func(ctx context.Context, filter bsoncore.Document) ([]bsoncore.Document, error)

KeyRetrieverFn is a callback used to retrieve keys from the key vault.

type LegacyOperationKind added in v1.1.0 

type LegacyOperationKind uint

LegacyOperationKind indicates if an operation is a legacy find, getMore, or killCursors. This is used in Operation.Execute, which will create legacy OP_QUERY, OP_GET_MORE, or OP_KILL_CURSORS instead of sending them as a command. 

const (
	LegacyNone LegacyOperationKind = iota
	LegacyFind
	LegacyGetMore
	LegacyKillCursors
	LegacyListCollections
	LegacyListIndexes
)

These constants represent the three different kinds of legacy operations.

type ListCollectionsBatchCursor added in v0.3.0 

type ListCollectionsBatchCursor struct {
	// contains filtered or unexported fields
}

ListCollectionsBatchCursor is a special batch cursor returned from ListCollections that properly handles current and legacy ListCollections operations.

func NewLegacyListCollectionsBatchCursor added in v0.3.0 

func NewLegacyListCollectionsBatchCursor(bc *BatchCursor) (*ListCollectionsBatchCursor, error)

NewLegacyListCollectionsBatchCursor creates a new legacy ListCollectionsCursor.

func NewListCollectionsBatchCursor added in v0.3.0 

func NewListCollectionsBatchCursor(bc *BatchCursor) (*ListCollectionsBatchCursor, error)

NewListCollectionsBatchCursor creates a new non-legacy ListCollectionsCursor.

func (*ListCollectionsBatchCursor) Batch added in v0.3.0 

func (lcbc *ListCollectionsBatchCursor) Batch() *bsoncore.DocumentSequence

Batch will return a DocumentSequence for the current batch of documents. The returned DocumentSequence is only valid until the next call to Next or Close.

func (*ListCollectionsBatchCursor) Close added in v0.3.0 

func (lcbc *ListCollectionsBatchCursor) Close(ctx context.Context) error

Close closes this batch cursor.

func (*ListCollectionsBatchCursor) Err added in v0.3.0 

func (lcbc *ListCollectionsBatchCursor) Err() error

Err returns the latest error encountered.

func (*ListCollectionsBatchCursor) ID added in v0.3.0 

func (lcbc *ListCollectionsBatchCursor) ID() int64

ID returns the cursor ID for this batch cursor.

func (*ListCollectionsBatchCursor) Next added in v0.3.0 

func (lcbc *ListCollectionsBatchCursor) Next(ctx context.Context) bool

Next indicates if there is another batch available. Returning false does not necessarily indicate that the cursor is closed. This method will return false when an empty batch is returned.

If Next returns true, there is a valid batch of documents available. If Next returns false, there is not a valid batch of documents available.

func (*ListCollectionsBatchCursor) Server added in v1.0.0 

func (lcbc *ListCollectionsBatchCursor) Server() Server

Server returns a pointer to the cursor's server.

type LocalAddresser added in v1.1.0 

type LocalAddresser interface {
	LocalAddress() address.Address
}

LocalAddresser is a type that is able to supply its local address

type MarkCommandFn added in v1.2.0 

type MarkCommandFn func(ctx context.Context, db string, cmd bsoncore.Document) (bsoncore.Document, error)

MarkCommandFn is a callback used to add encryption markings to a command.

type Operation added in v1.1.0 

type Operation struct {
	// CommandFn is used to create the command that will be wrapped in a wire message and sent to
	// the server. This function should only add the elements of the command and not start or end
	// the enclosing BSON document. Per the command API, the first element must be the name of the
	// command to run. This field is required.
	CommandFn func(dst []byte, desc description.SelectedServer) ([]byte, error)

	// Database is the database that the command will be run against. This field is required.
	Database string

	// Deployment is the MongoDB Deployment to use. While most of the time this will be multiple
	// servers, commands that need to run against a single, preselected server can use the
	// SingleServerDeployment type. Commands that need to run on a preselected connection can use
	// the SingleConnectionDeployment type.
	Deployment Deployment

	// ProcessResponseFn is called after a response to the command is returned. The server is
	// provided for types like Cursor that are required to run subsequent commands using the same
	// server.
	ProcessResponseFn func(response bsoncore.Document, srvr Server, desc description.Server) error

	// Selector is the server selector that's used during both initial server selection and
	// subsequent selection for retries. Depending on the Deployment implementation, the
	// SelectServer method may not actually be called.
	Selector description.ServerSelector

	// ReadPreference is the read preference that will be attached to the command. If this field is
	// not specified a default read preference of primary will be used.
	ReadPreference *readpref.ReadPref

	// ReadConcern is the read concern used when running read commands. This field should not be set
	// for write operations. If this field is set, it will be encoded onto the commands sent to the
	// server.
	ReadConcern *readconcern.ReadConcern

	// MinimumReadConcernWireVersion specifies the minimum wire version to add the read concern to
	// the command being executed.
	MinimumReadConcernWireVersion int32

	// WriteConcern is the write concern used when running write commands. This field should not be
	// set for read operations. If this field is set, it will be encoded onto the commands sent to
	// the server.
	WriteConcern *writeconcern.WriteConcern

	// MinimumWriteConcernWireVersion specifies the minimum wire version to add the write concern to
	// the command being executed.
	MinimumWriteConcernWireVersion int32

	// Client is the session used with this operation. This can be either an implicit or explicit
	// session. If the server selected does not support sessions and Client is specified the
	// behavior depends on the session type. If the session is implicit, the session fields will not
	// be encoded onto the command. If the session is explicit, an error will be returned. The
	// caller is responsible for ensuring that this field is nil if the Deployment does not support
	// sessions.
	Client *session.Client

	// Clock is a cluster clock, different from the one contained within a session.Client. This
	// allows updating cluster times for a global cluster clock while allowing individual session's
	// cluster clocks to be only updated as far as the last command that's been run.
	Clock *session.ClusterClock

	// RetryMode specifies how to retry. There are three modes that enable retry: RetryOnce,
	// RetryOncePerCommand, and RetryContext. For more information about what these modes do, please
	// refer to their definitions. Both RetryMode and Type must be set for retryability to be enabled.
	RetryMode *RetryMode

	// Type specifies the kind of operation this is. There is only one mode that enables retry: Write.
	// For more information about what this mode does, please refer to it's definition. Both Type and
	// RetryMode must be set for retryability to be enabled.
	Type Type

	// Batches contains the documents that are split when executing a write command that potentially
	// has more documents than can fit in a single command. This should only be specified for
	// commands that are batch compatible. For more information, please refer to the definition of
	// Batches.
	Batches *Batches

	// Legacy sets the legacy type for this operation. There are only 3 types that require legacy
	// support: find, getMore, and killCursors. For more information about LegacyOperationKind,
	// please refer to it's definition.
	Legacy LegacyOperationKind

	// CommandMonitor specifies the monitor to use for APM events. If this field is not set,
	// no events will be reported.
	CommandMonitor *event.CommandMonitor

	// Crypt specifies a Crypt object to use for automatic client side encryption and decryption.
	Crypt *Crypt
}

Operation is used to execute an operation. It contains all of the common code required to select a server, transform an operation into a command, write the command to a connection from the selected server, read a response from that connection, process the response, and potentially retry.

The required fields are Database, CommandFn, and Deployment. All other fields are optional.

While an Operation can be constructed manually, drivergen should be used to generate an implementation of an operation instead. This will ensure that there are helpers for constructing the operation and that this type isn't configured incorrectly.

func (Operation) Execute added in v1.1.0 

func (op Operation) Execute(ctx context.Context, scratch []byte) error

Execute runs this operation. The scratch parameter will be used and overwritten (potentially many times), this should mainly be used to enable pooling of byte slices.

func (Operation) Validate added in v1.1.0 

func (op Operation) Validate() error

Validate validates this operation, ensuring the fields are set properly.

type QueryFailureError added in v1.1.0 

type QueryFailureError struct {
	Message  string
	Response bsoncore.Document
}

QueryFailureError is an error representing a command failure as a document.

func (QueryFailureError) Error added in v1.1.0 

func (e QueryFailureError) Error() string

Error implements the error interface.

type ResponseError added in v1.1.0 

type ResponseError struct {
	Message string
	Wrapped error
}

ResponseError is an error parsing the response to a command.

func NewCommandResponseError added in v1.1.0 

func NewCommandResponseError(msg string, err error) ResponseError

NewCommandResponseError creates a CommandResponseError.

func (ResponseError) Error added in v1.1.0 

func (e ResponseError) Error() string

Error implements the error interface.

type RetryMode added in v1.1.0 

type RetryMode uint

RetryMode specifies the way that retries are handled for retryable operations. 

const (
	// RetryNone disables retrying.
	RetryNone RetryMode = iota
	// RetryOnce will enable retrying the entire operation once.
	RetryOnce
	// RetryOncePerCommand will enable retrying each command associated with an operation. For
	// example, if an insert is batch split into 4 commands then each of those commands is eligible
	// for one retry.
	RetryOncePerCommand
	// RetryContext will enable retrying until the context.Context's deadline is exceeded or it is
	// cancelled.
	RetryContext
)

These are the modes available for retrying.

func (RetryMode) Enabled added in v1.1.0 

func (rm RetryMode) Enabled() bool

Enabled returns if this RetryMode enables retrying.

type Server added in v1.1.0 

type Server interface {
	Connection(context.Context) (Connection, error)
}

Server represents a MongoDB server. Implementations should pool connections and handle the retrieving and returning of connections.

type SingleConnectionDeployment added in v1.1.0 

type SingleConnectionDeployment struct{ C Connection }

SingleConnectionDeployment is an implementation of Deployment that always returns the same Connection.

func (SingleConnectionDeployment) Connection added in v1.1.0 

func (ssd SingleConnectionDeployment) Connection(context.Context) (Connection, error)

Connection implements the Server interface. It always returns the embedded connection.

This method returns a Connection with a no-op Close method. This ensures that a SingleConnectionDeployment can be used across multiple operation executions.

func (SingleConnectionDeployment) Kind added in v1.1.0 

func (ssd SingleConnectionDeployment) Kind() description.TopologyKind

Kind implements the Deployment interface. It always returns description.Single.

func (SingleConnectionDeployment) SelectServer added in v1.1.0 

func (ssd SingleConnectionDeployment) SelectServer(context.Context, description.ServerSelector) (Server, error)

SelectServer implements the Deployment interface. This method does not use the description.SelectedServer provided and instead returns itself. The Connections returned from the Connection method have a no-op Close method.

func (SingleConnectionDeployment) SupportsRetryWrites added in v1.1.0 

func (ssd SingleConnectionDeployment) SupportsRetryWrites() bool

SupportsRetryWrites implements the Deployment interface. It always returns Type(0), because a single connection does not support retryability.

type SingleServerDeployment added in v1.1.0 

type SingleServerDeployment struct{ Server }

SingleServerDeployment is an implementation of Deployment that always returns a single server.

func (SingleServerDeployment) Kind added in v1.1.0 

func (SingleServerDeployment) Kind() description.TopologyKind

Kind implements the Deployment interface. It always returns description.Single.

func (SingleServerDeployment) SelectServer added in v1.1.0 

func (ssd SingleServerDeployment) SelectServer(context.Context, description.ServerSelector) (Server, error)

SelectServer implements the Deployment interface. This method does not use the description.SelectedServer provided and instead returns the embedded Server.

func (SingleServerDeployment) SupportsRetryWrites added in v1.1.0 

func (SingleServerDeployment) SupportsRetryWrites() bool

SupportsRetryWrites implements the Deployment interface. It always returns Type(0), because a single server does not support retryability.

type Subscriber added in v1.2.0 

type Subscriber interface {
	Subscribe() (*Subscription, error)
	Unsubscribe(*Subscription) error
}

Subscriber represents a type to which another type can subscribe. A subscription contains a channel that is updated with topology descriptions.

type Subscription added in v1.2.0 

type Subscription struct {
	Updates <-chan description.Topology
	ID      uint64
}

Subscription represents a subscription to topology updates. A subscriber can receive updates through the Updates field.

type Type added in v1.1.0 

type Type uint

Type specifies whether an operation is a read, write, or unknown. 

const (
	Write Type
	Read
)

THese are the availables types of Type.

type WriteCommandError added in v1.1.0 

type WriteCommandError struct {
	WriteConcernError *WriteConcernError
	WriteErrors       WriteErrors
}

WriteCommandError is an error for a write command.

func (WriteCommandError) Error added in v1.1.0 

func (wce WriteCommandError) Error() string

func (WriteCommandError) Retryable added in v1.1.0 

func (wce WriteCommandError) Retryable() bool

Retryable returns true if the error is retryable

func (WriteCommandError) UnsupportedStorageEngine added in v1.1.0 

func (wce WriteCommandError) UnsupportedStorageEngine() bool

UnsupportedStorageEngine returns whether or not the WriteCommandError comes from a retryable write being attempted against a server that has a storage engine where they are not supported

type WriteConcernError added in v1.1.0 

type WriteConcernError struct {
	Name    string
	Code    int64
	Message string
	Details bsoncore.Document
}

WriteConcernError is a write concern failure that occurred as a result of a write operation.

func (WriteConcernError) Error added in v1.1.0 

func (wce WriteConcernError) Error() string

func (WriteConcernError) NodeIsRecovering added in v1.1.0 

func (wce WriteConcernError) NodeIsRecovering() bool

NodeIsRecovering returns true if this error is a node is recovering error.

func (WriteConcernError) NodeIsShuttingDown added in v1.1.0 

func (wce WriteConcernError) NodeIsShuttingDown() bool

NodeIsShuttingDown returns true if this error is a node is shutting down error.

func (WriteConcernError) NotMaster added in v1.1.0 

func (wce WriteConcernError) NotMaster() bool

NotMaster returns true if this error is a not master error.

func (WriteConcernError) Retryable added in v1.1.0 

func (wce WriteConcernError) Retryable() bool

Retryable returns true if the error is retryable

type WriteError added in v1.1.0 

type WriteError struct {
	Index   int64
	Code    int64
	Message string
}

WriteError is a non-write concern failure that occurred as a result of a write operation.

func (WriteError) Error added in v1.1.0 

func (we WriteError) Error() string

type WriteErrors added in v1.1.0 

type WriteErrors []WriteError

WriteErrors is a group of non-write concern failures that occurred as a result of a write operation.

func (WriteErrors) Error added in v1.1.0 

func (we WriteErrors) Error() string

Source Files

View all Source files

Directories

Path Synopsis
Package auth is not for public use.
 Package auth is not for public use.
examples
cluster_monitoring
count
server_monitoring
workload
options
Package topology contains types that handles the discovery, monitoring, and selection of servers.
 Package topology contains types that handles the discovery, monitoring, and selection of servers.

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.