kv

package
v1.1.4 Latest Latest
Warning

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

Go to latest
Published: Jan 1, 2018 License: Apache-2.0 Imports: 38 Imported by: 0

Documentation

Overview

Package kv provides a key-value API to an underlying cockroach datastore. Cockroach itself provides a single, monolithic, sorted key value map, distributed over multiple nodes. Each node holds a set of key ranges. Package kv translates between the monolithic, logical map which Cockroach clients experience to the physically distributed key ranges which comprise the whole.

Package kv implements the logic necessary to locate appropriate nodes based on keys being read or written. In some cases, requests may span a range of keys, in which case multiple RPCs may be sent out.

Index

Constants

This section is empty.

Variables

This section is empty.

Functions

func InitSenderForLocalTestCluster

func InitSenderForLocalTestCluster(
	st *cluster.Settings,
	nodeDesc *roachpb.NodeDescriptor,
	tracer opentracing.Tracer,
	clock *hlc.Clock,
	latency time.Duration,
	stores client.Sender,
	stopper *stop.Stopper,
	gossip *gossip.Gossip,
) client.Sender

InitSenderForLocalTestCluster initializes a TxnCoordSender that can be used with LocalTestCluster.

Types

type BatchCall

type BatchCall struct {
	Reply *roachpb.BatchResponse
	Err   error
}

BatchCall contains a response and an RPC error (note that the response contains its own roachpb.Error, which is separate from BatchCall.Err), and is analogous to the net/rpc.Call struct.

type DBServer

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

A DBServer provides an HTTP server endpoint serving the key-value API. It accepts either JSON or serialized protobuf content types.

func NewDBServer

func NewDBServer(ctx *base.Config, sender client.Sender, stopper *stop.Stopper) *DBServer

NewDBServer allocates and returns a new DBServer.

func (*DBServer) Batch

func (s *DBServer) Batch(
	ctx context.Context, args *roachpb.BatchRequest,
) (br *roachpb.BatchResponse, err error)

Batch implements the roachpb.KVServer interface.

type DistSender

type DistSender struct {
	log.AmbientContext
	// contains filtered or unexported fields
}

A DistSender provides methods to access Cockroach's monolithic, distributed key value store. Each method invocation triggers a lookup or lookups to find replica metadata for implicated key ranges. RPCs are sent to one or more of the replicas to satisfy the method invocation.

func NewDistSender

func NewDistSender(cfg DistSenderConfig, g *gossip.Gossip) *DistSender

NewDistSender returns a batch.Sender instance which connects to the Cockroach cluster via the supplied gossip instance. Supplying a DistSenderContext or the fields within is optional. For omitted values, sane defaults will be used.

func (*DistSender) CountRanges

func (ds *DistSender) CountRanges(ctx context.Context, rs roachpb.RSpan) (int64, error)

CountRanges returns the number of ranges that encompass the given key span.

func (*DistSender) FirstRange

func (ds *DistSender) FirstRange() (*roachpb.RangeDescriptor, error)

FirstRange implements the RangeDescriptorDB interface. FirstRange returns the RangeDescriptor for the first range on the cluster, which is retrieved from the gossip protocol instead of the datastore.

func (*DistSender) GetParallelSendCount

func (ds *DistSender) GetParallelSendCount() int32

GetParallelSendCount returns the number of parallel batch requests the dist sender has dispatched in its lifetime.

func (*DistSender) LeaseHolderCache

func (ds *DistSender) LeaseHolderCache() *LeaseHolderCache

LeaseHolderCache gives access to the DistSender's lease cache.

func (*DistSender) Metrics

func (ds *DistSender) Metrics() DistSenderMetrics

Metrics returns a struct which contains metrics related to the distributed sender's activity.

func (*DistSender) RangeDescriptorCache

func (ds *DistSender) RangeDescriptorCache() *RangeDescriptorCache

RangeDescriptorCache gives access to the DistSender's range cache.

func (*DistSender) RangeLookup

func (ds *DistSender) RangeLookup(
	ctx context.Context, key roachpb.RKey, desc *roachpb.RangeDescriptor, useReverseScan bool,
) ([]roachpb.RangeDescriptor, []roachpb.RangeDescriptor, *roachpb.Error)

RangeLookup implements the RangeDescriptorDB interface. RangeLookup dispatches a RangeLookup request for the given metadata key to the replicas of the given range. Note that we allow inconsistent reads when doing range lookups for efficiency. Getting stale data is not a correctness problem but instead may infrequently result in additional latency as additional range lookups may be required. Note also that rangeLookup bypasses the DistSender's Send() method, so there is no error inspection and retry logic here; this is not an issue since the lookup performs a single inconsistent read only.

func (*DistSender) Send

Send implements the batch.Sender interface. It subdivides the Batch into batches admissible for sending (preventing certain illegal mixtures of requests), executes each individual part (which may span multiple ranges), and recombines the response.

When the request spans ranges, it is split by range and a partial subset of the batch request is sent to affected ranges in parallel.

The first write in a transaction may not arrive before writes to other ranges. This is relevant in the case of a BeginTransaction request. Intents written to other ranges before the transaction record is created will cause the transaction to abort early.

type DistSenderConfig

type DistSenderConfig struct {
	AmbientCtx log.AmbientContext

	Settings        *cluster.Settings
	Clock           *hlc.Clock
	RPCRetryOptions *retry.Options

	RPCContext        *rpc.Context
	RangeDescriptorDB RangeDescriptorDB

	TestingKnobs DistSenderTestingKnobs
	// contains filtered or unexported fields
}

DistSenderConfig holds configuration and auxiliary objects that can be passed to NewDistSender.

type DistSenderMetrics

type DistSenderMetrics struct {
	BatchCount             *metric.Counter
	PartialBatchCount      *metric.Counter
	SentCount              *metric.Counter
	LocalSentCount         *metric.Counter
	NextReplicaErrCount    *metric.Counter
	NotLeaseHolderErrCount *metric.Counter
	SlowRequestsCount      *metric.Gauge
}

DistSenderMetrics is the set of metrics for a given distributed sender.

type DistSenderTestingKnobs added in v1.1.0

type DistSenderTestingKnobs struct {
	// The RPC dispatcher. Defaults to grpc but can be changed here for
	// testing purposes.
	TransportFactory TransportFactory
}

DistSenderTestingKnobs is a part of the context used to control parts of the system.

func (*DistSenderTestingKnobs) ModuleTestingKnobs added in v1.1.0

func (*DistSenderTestingKnobs) ModuleTestingKnobs()

ModuleTestingKnobs is part of the base.ModuleTestingKnobs interface.

type EvictionToken

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

EvictionToken holds eviction state between calls to LookupRangeDescriptor.

func (*EvictionToken) Evict

func (et *EvictionToken) Evict(ctx context.Context) error

Evict instructs the EvictionToken to evict the RangeDescriptor it was created with from the rangeDescriptorCache.

func (*EvictionToken) EvictAndReplace

func (et *EvictionToken) EvictAndReplace(
	ctx context.Context, newDescs ...roachpb.RangeDescriptor,
) error

EvictAndReplace instructs the EvictionToken to evict the RangeDescriptor it was created with from the rangeDescriptorCache. It also allows the user to provide new RangeDescriptors to insert into the cache, all atomically. When called without arguments, EvictAndReplace will behave the same as Evict.

type LeaseHolderCache

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

A LeaseHolderCache is a cache of replica descriptors keyed by range ID.

func NewLeaseHolderCache

func NewLeaseHolderCache(size func() int64) *LeaseHolderCache

NewLeaseHolderCache creates a new leaseHolderCache of the given size. The underlying cache internally uses a hash map, so lookups are cheap.

func (*LeaseHolderCache) Lookup

Lookup returns the cached leader of the given range ID.

func (*LeaseHolderCache) Update

func (lc *LeaseHolderCache) Update(
	ctx context.Context, rangeID roachpb.RangeID, repDesc roachpb.ReplicaDescriptor,
)

Update invalidates the cached leader for the given range ID. If an empty replica descriptor is passed, the cached leader is evicted. Otherwise, the passed-in replica descriptor is cached.

type RangeDescriptorCache

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

RangeDescriptorCache is used to retrieve range descriptors for arbitrary keys. Descriptors are initially queried from storage using a RangeDescriptorDB, but are cached for subsequent lookups.

func NewRangeDescriptorCache

func NewRangeDescriptorCache(db RangeDescriptorDB, size func() int64) *RangeDescriptorCache

NewRangeDescriptorCache returns a new RangeDescriptorCache which uses the given RangeDescriptorDB as the underlying source of range descriptors.

func (*RangeDescriptorCache) EvictCachedRangeDescriptor

func (rdc *RangeDescriptorCache) EvictCachedRangeDescriptor(
	ctx context.Context, descKey roachpb.RKey, seenDesc *roachpb.RangeDescriptor, inclusive bool,
) error

EvictCachedRangeDescriptor will evict any cached range descriptors for the given key. It is intended that this method be called from a consumer of rangeDescriptorCache if the returned range descriptor is discovered to be stale. seenDesc should always be passed in and is used as the basis of a compare-and-evict (as pointers); if it is nil, eviction is unconditional but a warning will be logged.

func (*RangeDescriptorCache) GetCachedRangeDescriptor

func (rdc *RangeDescriptorCache) GetCachedRangeDescriptor(
	key roachpb.RKey, inclusive bool,
) (*roachpb.RangeDescriptor, error)

GetCachedRangeDescriptor retrieves the descriptor of the range which contains the given key. It returns nil if the descriptor is not found in the cache.

`inclusive` determines the behaviour at the range boundary: If set to true and `key` is the EndKey and StartKey of two adjacent ranges, the first range is returned instead of the second (which technically contains the given key).

func (*RangeDescriptorCache) InsertRangeDescriptors

func (rdc *RangeDescriptorCache) InsertRangeDescriptors(
	ctx context.Context, rs ...roachpb.RangeDescriptor,
) error

InsertRangeDescriptors inserts the provided descriptors in the cache. This is a no-op for the descriptors that are already present in the cache.

func (*RangeDescriptorCache) LookupRangeDescriptor

func (rdc *RangeDescriptorCache) LookupRangeDescriptor(
	ctx context.Context, key roachpb.RKey, evictToken *EvictionToken, useReverseScan bool,
) (*roachpb.RangeDescriptor, *EvictionToken, error)

LookupRangeDescriptor attempts to locate a descriptor for the range containing the given Key. This is done by first trying the cache, and then querying the two-level lookup table of range descriptors which cockroach maintains. The function should be provided with an EvictionToken if one was acquired from this function on a previous lookup. If not, an empty EvictionToken can be provided.

This method first looks up the specified key in the first level of range metadata, which returns the location of the key within the second level of range metadata. This second level location is then queried to retrieve a descriptor for the range where the key's value resides. Range descriptors retrieved during each search are cached for subsequent lookups.

This method returns the RangeDescriptor for the range containing the key's data and a token to manage evicting the RangeDescriptor if it is found to be stale, or an error if any occurred.

func (*RangeDescriptorCache) String

func (rdc *RangeDescriptorCache) String() string

type RangeDescriptorDB

type RangeDescriptorDB interface {
	// RangeLookup takes a meta key to look up descriptors for and a descriptor
	// of the range believed to hold the meta key. Two slices of range
	// descriptors are returned. The first of these slices holds descriptors
	// which contain the given key (possibly from intents), and the second holds
	// prefetched adjacent descriptors.
	RangeLookup(
		ctx context.Context,
		key roachpb.RKey,
		desc *roachpb.RangeDescriptor,
		useReverseScan bool,
	) ([]roachpb.RangeDescriptor, []roachpb.RangeDescriptor, *roachpb.Error)
	// FirstRange returns the descriptor for the first Range. This is the
	// Range containing all meta1 entries.
	FirstRange() (*roachpb.RangeDescriptor, error)
}

RangeDescriptorDB is a type which can query range descriptors from an underlying datastore. This interface is used by rangeDescriptorCache to initially retrieve information which will be cached.

type RangeIterator

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

A RangeIterator provides a mechanism for iterating over all ranges in a key span. A new RangeIterator must be positioned with Seek() to begin iteration.

RangeIterator is not thread-safe.

func NewRangeIterator

func NewRangeIterator(ds *DistSender) *RangeIterator

NewRangeIterator creates a new RangeIterator.

func (*RangeIterator) Desc

Desc returns the descriptor of the range at which the iterator is currently positioned. The iterator must be valid.

func (*RangeIterator) Error

func (ri *RangeIterator) Error() *roachpb.Error

Error returns the error the iterator encountered, if any. If the iterator has not been initialized, returns iterator error.

func (*RangeIterator) Key

func (ri *RangeIterator) Key() roachpb.RKey

Key returns the current key. The iterator must be valid.

func (*RangeIterator) LeaseHolder

func (ri *RangeIterator) LeaseHolder(ctx context.Context) (roachpb.ReplicaDescriptor, bool)

LeaseHolder returns the lease holder of the iterator's current range, if that information is present in the DistSender's LeaseHolderCache. The second return val is true if the descriptor has been found. The iterator must be valid.

func (*RangeIterator) NeedAnother

func (ri *RangeIterator) NeedAnother(rs roachpb.RSpan) bool

NeedAnother checks whether the iteration needs to continue to cover the remainder of the ranges described by the supplied key span. The iterator must be valid.

func (*RangeIterator) Next

func (ri *RangeIterator) Next(ctx context.Context)

Next advances the iterator to the next range. The direction of advance is dependent on whether the iterator is reversed. The iterator must be valid.

func (*RangeIterator) Seek

func (ri *RangeIterator) Seek(ctx context.Context, key roachpb.RKey, scanDir ScanDirection)

Seek positions the iterator at the specified key.

func (*RangeIterator) Token

func (ri *RangeIterator) Token() *EvictionToken

Token returns the eviction token corresponding to the range descriptor for the current iteration. The iterator must be valid.

func (*RangeIterator) Valid

func (ri *RangeIterator) Valid() bool

Valid returns whether the iterator is valid. To be valid, the iterator must be have been seeked to an initial position using Seek(), and must not have encountered an error.

type ReplicaInfo

type ReplicaInfo struct {
	roachpb.ReplicaDescriptor
	NodeDesc *roachpb.NodeDescriptor
}

ReplicaInfo extends the Replica structure with the associated node descriptor.

type ReplicaSlice

type ReplicaSlice []ReplicaInfo

A ReplicaSlice is a slice of ReplicaInfo.

func NewReplicaSlice

func NewReplicaSlice(gossip *gossip.Gossip, desc *roachpb.RangeDescriptor) ReplicaSlice

NewReplicaSlice creates a ReplicaSlice from the replicas listed in the range descriptor and using gossip to lookup node descriptors. Replicas on nodes that are not gossiped are omitted from the result.

func (ReplicaSlice) FindReplica

func (rs ReplicaSlice) FindReplica(storeID roachpb.StoreID) int

FindReplica returns the index of the replica which matches the specified store ID. If no replica matches, -1 is returned.

func (ReplicaSlice) FindReplicaByNodeID

func (rs ReplicaSlice) FindReplicaByNodeID(nodeID roachpb.NodeID) int

FindReplicaByNodeID returns the index of the replica which matches the specified node ID. If no replica matches, -1 is returned.

func (ReplicaSlice) Len

func (rs ReplicaSlice) Len() int

Len returns the total number of replicas in the slice.

func (ReplicaSlice) MoveToFront

func (rs ReplicaSlice) MoveToFront(i int)

MoveToFront moves the replica at the given index to the front of the slice, keeping the order of the remaining elements stable. The function will panic when invoked with an invalid index.

func (ReplicaSlice) OptimizeReplicaOrder

func (rs ReplicaSlice) OptimizeReplicaOrder(nodeDesc *roachpb.NodeDescriptor)

OptimizeReplicaOrder sorts the replicas in the order in which they're to be used for sending RPCs (meaning in the order in which they'll be probed for the lease). "Closer" (matching in more attributes) replicas are ordered first. If the current node is a replica, then it'll be the first one.

nodeDesc is the descriptor of the current node. It can be nil, in which case information about the current descriptor is not used in optimizing the order.

Note that this method is not concerned with any information the node might have about who the lease holder might be. If there is such info (e.g. in a LeaseHolderCache), the caller will probably want to further tweak the head of the ReplicaSlice.

func (ReplicaSlice) SortByCommonAttributePrefix

func (rs ReplicaSlice) SortByCommonAttributePrefix(attrs []string) int

SortByCommonAttributePrefix rearranges the ReplicaSlice by comparing the attributes to the given reference attributes. The basis for the comparison is that of the common prefix of replica attributes (i.e. the number of equal attributes, starting at the first), with a longer prefix sorting first. The number of attributes successfully matched to at least one replica is returned (hence, if the return value equals the length of the ReplicaSlice, at least one replica matched all attributes).

func (ReplicaSlice) Swap

func (rs ReplicaSlice) Swap(i, j int)

Swap swaps the replicas with indexes i and j.

type ScanDirection

type ScanDirection byte

ScanDirection determines the semantics of RangeIterator.Next() and RangeIterator.NeedAnother().

const (
	// Ascending means Next() will advance towards keys that compare higher.
	Ascending ScanDirection = iota
	// Descending means Next() will advance towards keys that compare lower.
	Descending
)

type SendOptions

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

A SendOptions structure describes the algorithm for sending RPCs to one or more replicas, depending on error conditions and how many successful responses are required.

type Transport

type Transport interface {
	// IsExhausted returns true if there are no more replicas to try.
	IsExhausted() bool

	// SendNext sends the rpc (captured at creation time) to the next
	// replica. May panic if the transport is exhausted. Should not
	// block; the transport is responsible for starting other goroutines
	// as needed.
	SendNext(context.Context, chan<- BatchCall)

	// NextReplica returns the replica descriptor of the replica to be tried in
	// the next call to SendNext. MoveToFront will cause the return value to
	// change. Returns a zero value if the transport is exhausted.
	NextReplica() roachpb.ReplicaDescriptor

	// MoveToFront locates the specified replica and moves it to the
	// front of the ordering of replicas to try. If the replica has
	// already been tried, it will be retried. If the specified replica
	// can't be found, this is a noop.
	MoveToFront(roachpb.ReplicaDescriptor)

	// Close is called when the transport is no longer needed. It may
	// cancel any pending RPCs without writing any response to the channel.
	Close()
}

Transport objects can send RPCs to one or more replicas of a range. All calls to Transport methods are made from a single thread, so Transports are not required to be thread-safe.

func GRPCTransportFactory added in v1.1.0

func GRPCTransportFactory(
	opts SendOptions, rpcContext *rpc.Context, replicas ReplicaSlice, args roachpb.BatchRequest,
) (Transport, error)

GRPCTransportFactory is the default TransportFactory, using GRPC.

type TransportFactory

type TransportFactory func(
	SendOptions, *rpc.Context, ReplicaSlice, roachpb.BatchRequest,
) (Transport, error)

TransportFactory encapsulates all interaction with the RPC subsystem, allowing it to be mocked out for testing. The factory function returns a Transport object which is used to send the given arguments to one or more replicas in the slice.

In addition to actually sending RPCs, the transport is responsible for ordering replicas in accordance with SendOptions.Ordering and transport-specific knowledge such as connection health or latency.

TODO(bdarnell): clean up this crufty interface; it was extracted verbatim from the non-abstracted code.

func SenderTransportFactory

func SenderTransportFactory(tracer opentracing.Tracer, sender client.Sender) TransportFactory

SenderTransportFactory wraps a client.Sender for use as a KV Transport. This is useful for tests that want to use DistSender without a full RPC stack.

type TxnCoordSender

type TxnCoordSender struct {
	log.AmbientContext
	// contains filtered or unexported fields
}

A TxnCoordSender is an implementation of client.Sender which wraps a lower-level Sender (either a storage.Stores or a DistSender) to which it sends commands. It acts as a man-in-the-middle, coordinating transaction state for clients. After a transaction is started, the TxnCoordSender starts asynchronously sending heartbeat messages to that transaction's txn record, to keep it live. It also keeps track of each written key or key range over the course of the transaction. When the transaction is committed or aborted, it clears accumulated write intents for the transaction.

func NewTxnCoordSender

func NewTxnCoordSender(
	ambient log.AmbientContext,
	st *cluster.Settings,
	wrapped client.Sender,
	clock *hlc.Clock,
	linearizable bool,
	stopper *stop.Stopper,
	txnMetrics TxnMetrics,
) *TxnCoordSender

NewTxnCoordSender creates a new TxnCoordSender for use from a KV distributed DB instance. ctx is the base context and is used for logs and traces when there isn't a more specific context available; it must have a Tracer set.

func (*TxnCoordSender) GetTxnState

func (tc *TxnCoordSender) GetTxnState(txnID uuid.UUID) (roachpb.Transaction, bool)

GetTxnState is part of the SenderWithDistSQLBackdoor interface.

func (*TxnCoordSender) Send

Send implements the batch.Sender interface. If the request is part of a transaction, the TxnCoordSender adds the transaction to a map of active transactions and begins heartbeating it. Every subsequent request for the same transaction updates the lastUpdate timestamp to prevent live transactions from being considered abandoned and garbage collected. Read/write mutating requests have their key or key range added to the transaction's interval tree of key ranges for eventual cleanup via resolved write intents; they're tagged to an outgoing EndTransaction request, with the receiving replica in charge of resolving them.

type TxnMetrics

type TxnMetrics struct {
	Aborts     *metric.CounterWithRates
	Commits    *metric.CounterWithRates
	Commits1PC *metric.CounterWithRates // Commits which finished in a single phase
	Abandons   *metric.CounterWithRates
	Durations  *metric.Histogram

	// Restarts is the number of times we had to restart the transaction.
	Restarts *metric.Histogram

	// Counts of restart types.
	RestartsWriteTooOld    *metric.Counter
	RestartsDeleteRange    *metric.Counter
	RestartsSerializable   *metric.Counter
	RestartsPossibleReplay *metric.Counter
}

TxnMetrics holds all metrics relating to KV transactions.

func MakeTxnMetrics

func MakeTxnMetrics(histogramWindow time.Duration) TxnMetrics

MakeTxnMetrics returns a TxnMetrics struct that contains metrics whose windowed portions retain data for approximately histogramWindow.

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL