foundationdb/fdbclient/DatabaseContext.h

413 lines
15 KiB
C++

/*
* DatabaseContext.h
*
* This source file is part of the FoundationDB open source project
*
* Copyright 2013-2018 Apple Inc. and the FoundationDB project authors
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#ifndef DatabaseContext_h
#define DatabaseContext_h
#include "flow/FastAlloc.h"
#include "flow/FastRef.h"
#include "fdbclient/StorageServerInterface.h"
#include "flow/genericactors.actor.h"
#include <vector>
#include <unordered_map>
#pragma once
#include "fdbclient/NativeAPI.actor.h"
#include "fdbclient/KeyRangeMap.h"
#include "fdbclient/CommitProxyInterface.h"
#include "fdbclient/SpecialKeySpace.actor.h"
#include "fdbrpc/QueueModel.h"
#include "fdbrpc/MultiInterface.h"
#include "flow/TDMetric.actor.h"
#include "fdbclient/EventTypes.actor.h"
#include "fdbrpc/ContinuousSample.h"
#include "fdbrpc/Smoother.h"
class StorageServerInfo : public ReferencedInterface<StorageServerInterface> {
public:
static Reference<StorageServerInfo> getInterface(DatabaseContext* cx,
StorageServerInterface const& interf,
LocalityData const& locality);
void notifyContextDestroyed();
~StorageServerInfo() override;
private:
DatabaseContext* cx;
StorageServerInfo(DatabaseContext* cx, StorageServerInterface const& interf, LocalityData const& locality)
: cx(cx), ReferencedInterface<StorageServerInterface>(interf, locality) {}
};
struct LocationInfo : MultiInterface<ReferencedInterface<StorageServerInterface>>, FastAllocated<LocationInfo> {
using Locations = MultiInterface<ReferencedInterface<StorageServerInterface>>;
explicit LocationInfo(const std::vector<Reference<ReferencedInterface<StorageServerInterface>>>& v)
: Locations(v) {}
LocationInfo(const std::vector<Reference<ReferencedInterface<StorageServerInterface>>>& v, bool hasCaches)
: Locations(v), hasCaches(hasCaches) {}
LocationInfo(const LocationInfo&) = delete;
LocationInfo(LocationInfo&&) = delete;
LocationInfo& operator=(const LocationInfo&) = delete;
LocationInfo& operator=(LocationInfo&&) = delete;
bool hasCaches = false;
Reference<Locations> locations() { return Reference<Locations>::addRef(this); }
};
using CommitProxyInfo = ModelInterface<CommitProxyInterface>;
using GrvProxyInfo = ModelInterface<GrvProxyInterface>;
class ClientTagThrottleData : NonCopyable {
private:
double tpsRate;
double expiration;
double lastCheck;
bool rateSet = false;
Smoother smoothRate;
Smoother smoothReleased;
public:
ClientTagThrottleData(ClientTagThrottleLimits const& limits)
: tpsRate(limits.tpsRate), expiration(limits.expiration), lastCheck(now()),
smoothRate(CLIENT_KNOBS->TAG_THROTTLE_SMOOTHING_WINDOW),
smoothReleased(CLIENT_KNOBS->TAG_THROTTLE_SMOOTHING_WINDOW) {
ASSERT(tpsRate >= 0);
smoothRate.reset(tpsRate);
}
void update(ClientTagThrottleLimits const& limits) {
ASSERT(limits.tpsRate >= 0);
this->tpsRate = limits.tpsRate;
if (!rateSet || expired()) {
rateSet = true;
smoothRate.reset(limits.tpsRate);
} else {
smoothRate.setTotal(limits.tpsRate);
}
expiration = limits.expiration;
}
void addReleased(int released) { smoothReleased.addDelta(released); }
bool expired() { return expiration <= now(); }
void updateChecked() { lastCheck = now(); }
bool canRecheck() { return lastCheck < now() - CLIENT_KNOBS->TAG_THROTTLE_RECHECK_INTERVAL; }
double throttleDuration() {
if (expiration <= now()) {
return 0.0;
}
double capacity =
(smoothRate.smoothTotal() - smoothReleased.smoothRate()) * CLIENT_KNOBS->TAG_THROTTLE_SMOOTHING_WINDOW;
if (capacity >= 1) {
return 0.0;
}
if (tpsRate == 0) {
return std::max(0.0, expiration - now());
}
return std::min(expiration - now(), capacity / tpsRate);
}
};
class WatchMetadata : public ReferenceCounted<WatchMetadata> {
public:
Key key;
Optional<Value> value;
Version version;
Promise<Version> watchPromise;
Future<Version> watchFuture;
Future<Void> watchFutureSS;
TransactionInfo info;
TagSet tags;
WatchMetadata(Key key, Optional<Value> value, Version version, TransactionInfo info, TagSet tags);
};
class DatabaseContext : public ReferenceCounted<DatabaseContext>, public FastAllocated<DatabaseContext>, NonCopyable {
public:
static DatabaseContext* allocateOnForeignThread() {
return (DatabaseContext*)DatabaseContext::operator new(sizeof(DatabaseContext));
}
// For internal (fdbserver) use only
static Database create(Reference<AsyncVar<ClientDBInfo>> clientInfo,
Future<Void> clientInfoMonitor,
LocalityData clientLocality,
bool enableLocalityLoadBalance,
TaskPriority taskID = TaskPriority::DefaultEndpoint,
bool lockAware = false,
int apiVersion = Database::API_VERSION_LATEST,
bool switchable = false);
~DatabaseContext();
Database clone() const {
return Database(new DatabaseContext(connectionFile,
clientInfo,
clientInfoMonitor,
taskID,
clientLocality,
enableLocalityLoadBalance,
lockAware,
internal,
apiVersion,
switchable));
}
std::pair<KeyRange, Reference<LocationInfo>> getCachedLocation(const KeyRef&, bool isBackward = false);
bool getCachedLocations(const KeyRangeRef&,
vector<std::pair<KeyRange, Reference<LocationInfo>>>&,
int limit,
bool reverse);
Reference<LocationInfo> setCachedLocation(const KeyRangeRef&, const vector<struct StorageServerInterface>&);
void invalidateCache(const KeyRef&, bool isBackward = false);
void invalidateCache(const KeyRangeRef&);
bool sampleReadTags() const;
bool sampleOnCost(uint64_t cost) const;
void updateProxies();
Reference<CommitProxyInfo> getCommitProxies(bool useProvisionalProxies);
Future<Reference<CommitProxyInfo>> getCommitProxiesFuture(bool useProvisionalProxies);
Reference<GrvProxyInfo> getGrvProxies(bool useProvisionalProxies);
Future<Void> onProxiesChanged();
Future<HealthMetrics> getHealthMetrics(bool detailed);
// Update the watch counter for the database
void addWatch();
void removeWatch();
// watch map operations
Reference<WatchMetadata> getWatchMetadata(KeyRef key) const;
KeyRef setWatchMetadata(Reference<WatchMetadata> metadata);
void deleteWatchMetadata(KeyRef key);
void clearWatchMetadata();
void setOption(FDBDatabaseOptions::Option option, Optional<StringRef> value);
Error deferredError;
bool lockAware;
bool isError() { return deferredError.code() != invalid_error_code; }
void checkDeferredError() {
if (isError()) {
throw deferredError;
}
}
int apiVersionAtLeast(int minVersion) { return apiVersion < 0 || apiVersion >= minVersion; }
Future<Void> onConnected(); // Returns after a majority of coordination servers are available and have reported a
// leader. The cluster file therefore is valid, but the database might be unavailable.
Reference<ClusterConnectionFile> getConnectionFile();
// Switch the database to use the new connection file, and recreate all pending watches for committed transactions.
//
// Meant to be used as part of a 'hot standby' solution to switch to the standby. A correct switch will involve
// advancing the version on the new cluster sufficiently far that any transaction begun with a read version from the
// old cluster will fail to commit. Assuming the above version-advancing is done properly, a call to
// switchConnectionFile guarantees that any read with a version from the old cluster will not be attempted on the
// new cluster.
Future<Void> switchConnectionFile(Reference<ClusterConnectionFile> standby);
Future<Void> connectionFileChanged();
bool switchable = false;
// Management API, Attempt to kill or suspend a process, return 1 for request sent out, 0 for failure
Future<int64_t> rebootWorker(StringRef address, bool check = false, int duration = 0);
// Management API, force the database to recover into DCID, causing the database to lose the most recently committed
// mutations
Future<Void> forceRecoveryWithDataLoss(StringRef dcId);
// Management API, create snapshot
Future<Void> createSnapshot(StringRef uid, StringRef snapshot_command);
// private:
explicit DatabaseContext(Reference<AsyncVar<Reference<ClusterConnectionFile>>> connectionFile,
Reference<AsyncVar<ClientDBInfo>> clientDBInfo,
Future<Void> clientInfoMonitor,
TaskPriority taskID,
LocalityData const& clientLocality,
bool enableLocalityLoadBalance,
bool lockAware,
bool internal = true,
int apiVersion = Database::API_VERSION_LATEST,
bool switchable = false);
explicit DatabaseContext(const Error& err);
void expireThrottles();
// Key DB-specific information
Reference<AsyncVar<Reference<ClusterConnectionFile>>> connectionFile;
AsyncTrigger proxiesChangeTrigger;
Future<Void> monitorProxiesInfoChange;
Reference<CommitProxyInfo> commitProxies;
Reference<GrvProxyInfo> grvProxies;
bool proxyProvisional; // Provisional commit proxy and grv proxy are used at the same time.
UID proxiesLastChange;
LocalityData clientLocality;
QueueModel queueModel;
bool enableLocalityLoadBalance;
struct VersionRequest {
SpanID spanContext;
Promise<GetReadVersionReply> reply;
TagSet tags;
Optional<UID> debugID;
VersionRequest(SpanID spanContext, TagSet tags = TagSet(), Optional<UID> debugID = Optional<UID>())
: spanContext(spanContext), tags(tags), debugID(debugID) {}
};
// Transaction start request batching
struct VersionBatcher {
PromiseStream<VersionRequest> stream;
Future<Void> actor;
};
std::map<uint32_t, VersionBatcher> versionBatcher;
AsyncTrigger connectionFileChangedTrigger;
// Disallow any reads at a read version lower than minAcceptableReadVersion. This way the client does not have to
// trust that the read version (possibly set manually by the application) is actually from the correct cluster.
// Updated everytime we get a GRV response
Version minAcceptableReadVersion = std::numeric_limits<Version>::max();
void validateVersion(Version);
// Client status updater
struct ClientStatusUpdater {
std::vector<std::pair<std::string, BinaryWriter>> inStatusQ;
std::vector<std::pair<std::string, BinaryWriter>> outStatusQ;
Future<Void> actor;
};
ClientStatusUpdater clientStatusUpdater;
// Cache of location information
int locationCacheSize;
CoalescedKeyRangeMap<Reference<LocationInfo>> locationCache;
std::map<UID, StorageServerInfo*> server_interf;
UID dbId;
bool internal; // Only contexts created through the C client and fdbcli are non-internal
PrioritizedTransactionTagMap<ClientTagThrottleData> throttledTags;
CounterCollection cc;
Counter transactionReadVersions;
Counter transactionReadVersionsThrottled;
Counter transactionReadVersionsCompleted;
Counter transactionReadVersionBatches;
Counter transactionBatchReadVersions;
Counter transactionDefaultReadVersions;
Counter transactionImmediateReadVersions;
Counter transactionBatchReadVersionsCompleted;
Counter transactionDefaultReadVersionsCompleted;
Counter transactionImmediateReadVersionsCompleted;
Counter transactionLogicalReads;
Counter transactionPhysicalReads;
Counter transactionPhysicalReadsCompleted;
Counter transactionGetKeyRequests;
Counter transactionGetValueRequests;
Counter transactionGetRangeRequests;
Counter transactionWatchRequests;
Counter transactionGetAddressesForKeyRequests;
Counter transactionBytesRead;
Counter transactionKeysRead;
Counter transactionMetadataVersionReads;
Counter transactionCommittedMutations;
Counter transactionCommittedMutationBytes;
Counter transactionSetMutations;
Counter transactionClearMutations;
Counter transactionAtomicMutations;
Counter transactionsCommitStarted;
Counter transactionsCommitCompleted;
Counter transactionKeyServerLocationRequests;
Counter transactionKeyServerLocationRequestsCompleted;
Counter transactionStatusRequests;
Counter transactionsTooOld;
Counter transactionsFutureVersions;
Counter transactionsNotCommitted;
Counter transactionsMaybeCommitted;
Counter transactionsResourceConstrained;
Counter transactionsProcessBehind;
Counter transactionsThrottled;
Counter transactionsExpensiveClearCostEstCount;
ContinuousSample<double> latencies, readLatencies, commitLatencies, GRVLatencies, mutationsPerCommit,
bytesPerCommit;
int outstandingWatches;
int maxOutstandingWatches;
int snapshotRywEnabled;
int transactionTracingEnabled;
Future<Void> logger;
Future<Void> throttleExpirer;
TaskPriority taskID;
Int64MetricHandle getValueSubmitted;
EventMetricHandle<GetValueComplete> getValueCompleted;
Reference<AsyncVar<ClientDBInfo>> clientInfo;
Future<Void> clientInfoMonitor;
Future<Void> connected;
Reference<AsyncVar<Optional<ClusterInterface>>> statusClusterInterface;
Future<Void> statusLeaderMon;
double lastStatusFetch;
int apiVersion;
int mvCacheInsertLocation;
std::vector<std::pair<Version, Optional<Value>>> metadataVersionCache;
HealthMetrics healthMetrics;
double healthMetricsLastUpdated;
double detailedHealthMetricsLastUpdated;
Smoother smoothMidShardSize;
UniqueOrderedOptionList<FDBTransactionOptions> transactionDefaults;
Future<Void> cacheListMonitor;
AsyncTrigger updateCache;
std::vector<std::unique_ptr<SpecialKeyRangeReadImpl>> specialKeySpaceModules;
std::unique_ptr<SpecialKeySpace> specialKeySpace;
void registerSpecialKeySpaceModule(SpecialKeySpace::MODULE module,
SpecialKeySpace::IMPLTYPE type,
std::unique_ptr<SpecialKeyRangeReadImpl>&& impl);
static bool debugUseTags;
static const std::vector<std::string> debugTransactionTagChoices;
std::unordered_map<KeyRef, Reference<WatchMetadata>> watchMap;
};
#endif