wbtiger
/
foundationdb
mirror of https://github.com/apple/foundationdb.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
foundationdb/fdbserver/workloads/Serializability.actor.cpp

511 lines
18 KiB
C++
Raw Normal View History

Initial repository commit
2017-05-26 04:48:44 +08:00
/*
* Serializability.actor.cpp
*
* This source file is part of the FoundationDB open source project
*
* Copyright 2013-2018 Apple Inc. and the FoundationDB project authors
remove trailing whitespace from our copyright headers ; fixed formatting of python setup.py
2018-02-22 02:25:11 +08:00
*
Initial repository commit
2017-05-26 04:48:44 +08:00
* 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
remove trailing whitespace from our copyright headers ; fixed formatting of python setup.py
2018-02-22 02:25:11 +08:00
*
Initial repository commit
2017-05-26 04:48:44 +08:00
* http://www.apache.org/licenses/LICENSE-2.0
remove trailing whitespace from our copyright headers ; fixed formatting of python setup.py
2018-02-22 02:25:11 +08:00
*
Initial repository commit
2017-05-26 04:48:44 +08:00
* 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.
*/
Use proper fwd decl in NativeAPI Also NativeAPI.h -> NativeAPI.actor.h
2019-02-18 07:41:16 +08:00
#include "fdbclient/NativeAPI.actor.h"
Use correct fwd decls in TesterInterface Also TesterInterface.h -> TesterInterface.actor.h
2019-02-18 11:25:16 +08:00
#include "fdbserver/TesterInterface.actor.h"
Initial repository commit
2017-05-26 04:48:44 +08:00
#include "fdbclient/ReadYourWrites.h"
#include "flow/ActorCollection.h"
Use fwd decls in workloads Also workloads.h -> workloads.actor.h
2019-02-18 11:18:30 +08:00
#include "fdbserver/workloads/workloads.actor.h"
Rewrite all files to have #include actorcompiler.h as the last include.
2018-08-11 06:18:24 +08:00
#include "flow/actorcompiler.h" // This must be the last #include.
Initial repository commit
2017-05-26 04:48:44 +08:00
struct SerializabilityWorkload : TestWorkload {
double testDuration;
bool adjacentKeys;
int nodes;
int numOps;
std::pair<int,int> valueSizeRange;
int maxClearSize;
std::string keyPrefix;
bool success;
struct GetRangeOperation {
KeySelector begin;
KeySelector end;
int limit;
bool snapshot;
bool reverse;
};
struct GetKeyOperation {
KeySelector key;
bool snapshot;
};
struct GetOperation {
Key key;
bool snapshot;
};
struct TransactionOperation {
Optional<Standalone<MutationRef>> mutationOp;
Optional<GetRangeOperation> getRangeOp;
Optional<GetKeyOperation> getKeyOp;
Optional<GetOperation> getOp;
Optional<Key> watchOp;
Optional<KeyRange> writeConflictOp;
Optional<KeyRange> readConflictOp;
};
SerializabilityWorkload(WorkloadContext const& wcx)
: TestWorkload(wcx), success(true) {
testDuration = getOption( options, LiteralStringRef("testDuration"), 30.0 );
numOps = getOption( options, LiteralStringRef("numOps"), 21 );
nodes = getOption( options, LiteralStringRef("nodes"), 1000 );
adjacentKeys = false; //g_random->random01() < 0.5;
valueSizeRange = std::make_pair( 0, 100 );
//keyPrefix = "\x02";
maxClearSize = g_random->randomInt(10, 2*nodes);
if( clientId == 0 )
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent("SerializabilityConfiguration").detail("Nodes", nodes).detail("AdjacentKeys", adjacentKeys).detail("ValueSizeMin", valueSizeRange.first).detail("ValueSizeMax", valueSizeRange.second).detail("MaxClearSize", maxClearSize);
Initial repository commit
2017-05-26 04:48:44 +08:00
}
virtual std::string description() { return "Serializability"; }
virtual Future<Void> setup( Database const& cx ) {
return Void();
}
virtual Future<Void> start( Database const& cx ) {
if( clientId == 0 )
return _start( cx, this );
return Void();
}
virtual Future<bool> check( Database const& cx ) {
return success;
}
virtual void getMetrics( vector<PerfMetric>& m ) {
}
Value getRandomValue() {
return Value( std::string( g_random->randomInt(valueSizeRange.first,valueSizeRange.second+1), 'x' ) );
}
Key getRandomKey() {
return getKeyForIndex( g_random->randomInt(0, nodes ) );
}
Key getKeyForIndex( int idx ) {
if( adjacentKeys ) {
return Key( idx ? keyPrefix + std::string( idx, '\x00' ) : "" );
} else {
return Key( keyPrefix + format( "%010d", idx ) );
}
}
KeySelector getRandomKeySelector() {
int scale = 1 << g_random->randomInt(0,14);
return KeySelectorRef( getRandomKey(), g_random->random01() < 0.5, g_random->randomInt(-scale, scale) );
}
KeyRange getRandomRange(int sizeLimit) {
int startLocation = g_random->randomInt(0, nodes);
int scale = g_random->randomInt(0, g_random->randomInt(2, 5) * g_random->randomInt(2, 5));
int endLocation = startLocation + g_random->randomInt(0, 1+std::min(sizeLimit, std::min(nodes-startLocation, 1<<scale)));
return KeyRangeRef( getKeyForIndex( startLocation ), getKeyForIndex( endLocation ) );
}
std::vector<TransactionOperation> randomTransaction() {
int maxOps = g_random->randomInt( 1, numOps );
std::vector<TransactionOperation> result;
bool hasMutation = false;
for(int j = 0; j < maxOps; j++ ) {
int operationType = g_random->randomInt(0, 20);
TransactionOperation op;
if( operationType == 0 ) {
GetKeyOperation getKey;
getKey.key = getRandomKeySelector();
getKey.snapshot = g_random->random01() < 0.5;
op.getKeyOp = getKey;
} else if( operationType == 1 ) {
GetRangeOperation getRange;
getRange.begin = getRandomKeySelector();
getRange.end = getRandomKeySelector();
getRange.limit = g_random->randomInt(0, 1<<g_random->randomInt(1, 10));
getRange.reverse = g_random->random01() < 0.5;
getRange.snapshot = g_random->random01() < 0.5;
op.getRangeOp = getRange;
} else if( operationType == 2 ) {
GetOperation getOp;
getOp.key = getRandomKey();
getOp.snapshot = g_random->random01() < 0.5;
op.getOp = getOp;
} else if( operationType == 3 ) {
KeyRange range = getRandomRange( maxClearSize );
op.mutationOp = MutationRef(MutationRef::ClearRange, range.begin, range.end);
if(!range.empty())
hasMutation = true;
} else if( operationType == 4 ) {
KeyRange range = singleKeyRange(getRandomKey());
op.mutationOp = MutationRef(MutationRef::ClearRange, range.begin, range.end);
hasMutation = true;
} else if( operationType == 5 ) {
op.watchOp = getRandomKey();
} else if( operationType == 6 ) {
op.writeConflictOp = getRandomRange( maxClearSize );
} else if( operationType == 7 ) {
op.readConflictOp = getRandomRange( maxClearSize );
} else if( operationType == 8 ) {
Key key = getRandomKey();
Value value = getRandomValue();
MutationRef::Type opType;
Modified existing behavior of MIN and AND atomic ops. The new behavior results in a 'SET' if the atomic op is performed on a non -existing key. Added new atomic ops ByteMin and ByteMax that does lexicographic comparison of byte strings.
2017-10-11 04:02:22 +08:00
switch( g_random->randomInt(0,8) ) {
Initial repository commit
2017-05-26 04:48:44 +08:00
case 0:
opType = MutationRef::AddValue;
break;
case 1:
opType = MutationRef::And;
break;
case 2:
opType = MutationRef::Or;
break;
case 3:
opType = MutationRef::Xor;
break;
case 4:
opType = MutationRef::Max;
break;
case 5:
opType = MutationRef::Min;
break;
Modified existing behavior of MIN and AND atomic ops. The new behavior results in a 'SET' if the atomic op is performed on a non -existing key. Added new atomic ops ByteMin and ByteMax that does lexicographic comparison of byte strings.
2017-10-11 04:02:22 +08:00
case 6:
opType = MutationRef::ByteMin;
break;
case 7:
opType = MutationRef::ByteMax;
break;
Initial repository commit
2017-05-26 04:48:44 +08:00
}
op.mutationOp = MutationRef(opType, key, value);
hasMutation = true;
} else if( operationType >= 9 ) {
Key key = getRandomKey();
Value value = getRandomValue();
op.mutationOp = MutationRef(MutationRef::SetValue, key, value);
hasMutation = true;
}
result.push_back(op);
}
if(!hasMutation) {
Key key = getRandomKey();
Value value = getRandomValue();
TransactionOperation op;
op.mutationOp = MutationRef(MutationRef::SetValue, key, value);
result.push_back(op);
}
return result;
}
template <class T>
static void dontCheck( std::vector<Future<T>>& futures ) {
// Replace the last future in the vector with one that will be completed at the same time and
// with the same error status, but has a constant result. This is used to suppress the results
// of reads that aren't deterministic in the test context.
futures.back() = tag(::success(futures.back()), T());
}
ACTOR static Future<Void> runTransaction( ReadYourWritesTransaction* tr,
std::vector<TransactionOperation> ops,
std::vector<Future<Optional<Value>>>* getFutures,
std::vector<Future<Key>>* getKeyFutures,
std::vector<Future<Standalone<RangeResultRef>>>* getRangeFutures,
std::vector<Future<Void>>* watchFutures,
bool checkSnapshotReads)
{
state int opNum = 0;
for(; opNum < ops.size(); opNum++) {
if(ops[opNum].getKeyOp.present()) {
auto& op = ops[opNum].getKeyOp.get();
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
//TraceEvent("SRL_GetKey").detail("Key", op.key.toString()).detail("Snapshot", op.snapshot);
Initial repository commit
2017-05-26 04:48:44 +08:00
getKeyFutures->push_back(tr->getKey(op.key, op.snapshot));
if (op.snapshot && !checkSnapshotReads)
dontCheck(*getKeyFutures);
} else if(ops[opNum].getOp.present()) {
auto& op = ops[opNum].getOp.get();
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
//TraceEvent("SRL_Get").detail("Key", printable(op.key)).detail("Snapshot", op.snapshot);
Initial repository commit
2017-05-26 04:48:44 +08:00
getFutures->push_back(tr->get(op.key, op.snapshot));
if (op.snapshot && !checkSnapshotReads)
dontCheck(*getFutures);
} else if(ops[opNum].getRangeOp.present()) {
auto& op = ops[opNum].getRangeOp.get();
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
//TraceEvent("SRL_GetRange").detail("Begin", op.begin.toString()).detail("End", op.end.toString()).detail("Limit", op.limit).detail("Snapshot", op.snapshot).detail("Reverse", op.reverse);
Initial repository commit
2017-05-26 04:48:44 +08:00
getRangeFutures->push_back(tr->getRange(op.begin, op.end, op.limit, op.snapshot, op.reverse));
if (op.snapshot && !checkSnapshotReads)
dontCheck(*getRangeFutures);
} else if(ops[opNum].mutationOp.present()) {
auto& op = ops[opNum].mutationOp.get();
if(op.type == MutationRef::SetValue) {
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
//TraceEvent("SRL_Set").detail("Mutation", op.toString());
Initial repository commit
2017-05-26 04:48:44 +08:00
tr->set(op.param1, op.param2);
} else if(op.type == MutationRef::ClearRange) {
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
//TraceEvent("SRL_Clear").detail("Mutation", op.toString());
Initial repository commit
2017-05-26 04:48:44 +08:00
tr->clear(KeyRangeRef(op.param1, op.param2));
} else {
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
//TraceEvent("SRL_AtomicOp").detail("Mutation", op.toString());
Initial repository commit
2017-05-26 04:48:44 +08:00
tr->atomicOp(op.param1, op.param2, op.type);
}
} else if(ops[opNum].readConflictOp.present()) {
auto& op = ops[opNum].readConflictOp.get();
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
//TraceEvent("SRL_ReadConflict").detail("Range", printable(op));
Initial repository commit
2017-05-26 04:48:44 +08:00
tr->addReadConflictRange(op);
} else if(ops[opNum].watchOp.present()) {
auto& op = ops[opNum].watchOp.get();
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
//TraceEvent("SRL_Watch").detail("Key", printable(op));
Initial repository commit
2017-05-26 04:48:44 +08:00
watchFutures->push_back(tr->watch(op));
} else if(ops[opNum].writeConflictOp.present()) {
auto& op = ops[opNum].writeConflictOp.get();
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
//TraceEvent("SRL_WriteConflict").detail("Range", printable(op));
Initial repository commit
2017-05-26 04:48:44 +08:00
tr->addWriteConflictRange(op);
}
//sometimes wait for a random operation
if( g_random->random01() < 0.2 ) {
state int waitType = g_random->randomInt(0, 4);
loop {
if(waitType == 0 && getFutures->size()) {
Rewrite all `Void _ = wait(...)` -> `wait(...)`. This takes advantage of the new actorcompiler functionality to avoid having duplicate definitions of `Void _` when trying to feed the un-actorompiled source through clang.
2018-08-11 04:57:10 +08:00
wait( ::success( g_random->randomChoice(*getFutures) ) );
Initial repository commit
2017-05-26 04:48:44 +08:00
break;
} else if(waitType == 1 && getKeyFutures->size()) {
Rewrite all `Void _ = wait(...)` -> `wait(...)`. This takes advantage of the new actorcompiler functionality to avoid having duplicate definitions of `Void _` when trying to feed the un-actorompiled source through clang.
2018-08-11 04:57:10 +08:00
wait( ::success( g_random->randomChoice(*getKeyFutures) ) );
Initial repository commit
2017-05-26 04:48:44 +08:00
break;
} else if(waitType == 2 && getRangeFutures->size()) {
Rewrite all `Void _ = wait(...)` -> `wait(...)`. This takes advantage of the new actorcompiler functionality to avoid having duplicate definitions of `Void _` when trying to feed the un-actorompiled source through clang.
2018-08-11 04:57:10 +08:00
wait( ::success( g_random->randomChoice(*getRangeFutures) ) );
Initial repository commit
2017-05-26 04:48:44 +08:00
break;
} else if(waitType == 3) {
Rewrite all `Void _ = wait(...)` -> `wait(...)`. This takes advantage of the new actorcompiler functionality to avoid having duplicate definitions of `Void _` when trying to feed the un-actorompiled source through clang.
2018-08-11 04:57:10 +08:00
wait( delay(0.001*g_random->random01()));
Initial repository commit
2017-05-26 04:48:44 +08:00
break;
}
waitType = (waitType + 1) % 4;
}
}
}
return Void();
}
ACTOR static Future<Standalone<RangeResultRef>> getDatabaseContents( Database cx, int nodes ) {
state ReadYourWritesTransaction tr(cx);
Standalone<RangeResultRef> result = wait( tr.getRange(normalKeys, nodes+1) );
ASSERT(result.size() <= nodes);
return result;
}
ACTOR static Future<Void> resetDatabase( Database cx, Standalone<VectorRef<KeyValueRef>> data ) {
state ReadYourWritesTransaction tr(cx);
tr.clear(normalKeys);
for(auto kv : data)
tr.set(kv.key, kv.value);
Rewrite all `Void _ = wait(...)` -> `wait(...)`. This takes advantage of the new actorcompiler functionality to avoid having duplicate definitions of `Void _` when trying to feed the un-actorompiled source through clang.
2018-08-11 04:57:10 +08:00
wait( tr.commit() );
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
//TraceEvent("SRL_Reset");
Initial repository commit
2017-05-26 04:48:44 +08:00
return Void();
}
ACTOR Future<Void> _start( Database cx, SerializabilityWorkload* self ) {
state double startTime = now();
loop {
state std::vector<ReadYourWritesTransaction> tr;
state std::vector<std::vector<Future<Optional<Value>>>> getFutures;
state std::vector<std::vector<Future<Key>>> getKeyFutures;
state std::vector<std::vector<Future<Standalone<RangeResultRef>>>> getRangeFutures;
state std::vector<std::vector<Future<Void>>> watchFutures;
for(int i = 0; i < 5; i++) {
tr.push_back(ReadYourWritesTransaction(cx));
getFutures.push_back(std::vector<Future<Optional<Value>>>());
getKeyFutures.push_back(std::vector<Future<Key>>());
getRangeFutures.push_back(std::vector<Future<Standalone<RangeResultRef>>>());
watchFutures.push_back(std::vector<Future<Void>>());
}
try {
if(now() - startTime > self->testDuration)
return Void();
//Generate initial data
state Standalone<VectorRef<KeyValueRef>> initialData;
int initialAmount = g_random->randomInt(0, 100);
for(int i = 0; i < initialAmount; i++) {
Key key = self->getRandomKey();
Value value = self->getRandomValue();
initialData.push_back_deep(initialData.arena(), KeyValueRef(key, value));
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
//TraceEvent("SRL_Init").detail("Key", printable(key)).detail("Value", printable(value));
Initial repository commit
2017-05-26 04:48:44 +08:00
}
//Generate three random transactions
state std::vector<TransactionOperation> a = self->randomTransaction();
state std::vector<TransactionOperation> b = self->randomTransaction();
state std::vector<TransactionOperation> c = self->randomTransaction();
//reset database to known state
Rewrite all `Void _ = wait(...)` -> `wait(...)`. This takes advantage of the new actorcompiler functionality to avoid having duplicate definitions of `Void _` when trying to feed the un-actorompiled source through clang.
2018-08-11 04:57:10 +08:00
wait( resetDatabase(cx, initialData) );
Initial repository commit
2017-05-26 04:48:44 +08:00
Rewrite all `Void _ = wait(...)` -> `wait(...)`. This takes advantage of the new actorcompiler functionality to avoid having duplicate definitions of `Void _` when trying to feed the un-actorompiled source through clang.
2018-08-11 04:57:10 +08:00
wait( runTransaction(&tr[0], a, &getFutures[0], &getKeyFutures[0], &getRangeFutures[0], &watchFutures[0], true) );
wait( tr[0].commit() );
Initial repository commit
2017-05-26 04:48:44 +08:00
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
//TraceEvent("SRL_FinishedA");
Initial repository commit
2017-05-26 04:48:44 +08:00
Rewrite all `Void _ = wait(...)` -> `wait(...)`. This takes advantage of the new actorcompiler functionality to avoid having duplicate definitions of `Void _` when trying to feed the un-actorompiled source through clang.
2018-08-11 04:57:10 +08:00
wait( runTransaction(&tr[1], b, &getFutures[0], &getKeyFutures[0], &getRangeFutures[0], &watchFutures[0], true) );
wait( tr[1].commit() );
Initial repository commit
2017-05-26 04:48:44 +08:00
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
//TraceEvent("SRL_FinishedB");
Initial repository commit
2017-05-26 04:48:44 +08:00
Rewrite all `Void _ = wait(...)` -> `wait(...)`. This takes advantage of the new actorcompiler functionality to avoid having duplicate definitions of `Void _` when trying to feed the un-actorompiled source through clang.
2018-08-11 04:57:10 +08:00
wait( runTransaction(&tr[2], c, &getFutures[2], &getKeyFutures[2], &getRangeFutures[2], &watchFutures[2], false) );
wait( tr[2].commit() );
Initial repository commit
2017-05-26 04:48:44 +08:00
//get contents of database
state Standalone<RangeResultRef> result1 = wait( getDatabaseContents(cx, self->nodes) );
//reset database to known state
Rewrite all `Void _ = wait(...)` -> `wait(...)`. This takes advantage of the new actorcompiler functionality to avoid having duplicate definitions of `Void _` when trying to feed the un-actorompiled source through clang.
2018-08-11 04:57:10 +08:00
wait( resetDatabase(cx, initialData) );
Initial repository commit
2017-05-26 04:48:44 +08:00
Rewrite all `Void _ = wait(...)` -> `wait(...)`. This takes advantage of the new actorcompiler functionality to avoid having duplicate definitions of `Void _` when trying to feed the un-actorompiled source through clang.
2018-08-11 04:57:10 +08:00
wait( runTransaction(&tr[3], a, &getFutures[3], &getKeyFutures[3], &getRangeFutures[3], &watchFutures[3], true) );
wait( runTransaction(&tr[3], b, &getFutures[3], &getKeyFutures[3], &getRangeFutures[3], &watchFutures[3], true) );
wait( runTransaction(&tr[4], c, &getFutures[4], &getKeyFutures[4], &getRangeFutures[4], &watchFutures[4], false) );
wait( tr[3].commit() );
wait( tr[4].commit() );
Initial repository commit
2017-05-26 04:48:44 +08:00
//get contents of database
Standalone<RangeResultRef> result2 = wait( getDatabaseContents(cx, self->nodes) );
if(result1.size() != result2.size()) {
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent(SevError, "SRL_ResultMismatch").detail("Size1", result1.size()).detail("Size2", result2.size());
Initial repository commit
2017-05-26 04:48:44 +08:00
for(auto kv : result1)
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent("SRL_Result1").detail("Kv", printable(kv));
Initial repository commit
2017-05-26 04:48:44 +08:00
for(auto kv : result2)
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent("SRL_Result2").detail("Kv", printable(kv));
Initial repository commit
2017-05-26 04:48:44 +08:00
ASSERT(false);
}
for(int i = 0; i < result1.size(); i++) {
if(result1[i] != result2[i]) {
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent(SevError, "SRL_ResultMismatch").detail("I", i).detail("Result1", printable(result1[i])).detail("Result2", printable(result2[i]))
.detail("Result1Value", printable(result1[i].value)).detail("Result2Value", printable(result2[i].value));
Initial repository commit
2017-05-26 04:48:44 +08:00
for(auto kv : result1)
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent("SRL_Result1").detail("Kv", printable(kv));
Initial repository commit
2017-05-26 04:48:44 +08:00
for(auto kv : result2)
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent("SRL_Result2").detail("Kv", printable(kv));
Initial repository commit
2017-05-26 04:48:44 +08:00
ASSERT(false);
}
}
for(int i = 0; i < getFutures[0].size(); i++) {
ASSERT(getFutures[0][i].get() == getFutures[3][i].get());
}
for(int i = 0; i < getFutures[1].size(); i++) {
ASSERT(getFutures[1][i].get() == getFutures[3][getFutures[0].size()+i].get());
}
for(int i = 0; i < getFutures[2].size(); i++) {
ASSERT(getFutures[2][i].get() == getFutures[4][i].get());
}
for(int i = 0; i < getKeyFutures[0].size(); i++) {
ASSERT(getKeyFutures[0][i].get() == getKeyFutures[3][i].get());
}
for(int i = 0; i < getKeyFutures[1].size(); i++) {
ASSERT(getKeyFutures[1][i].get() == getKeyFutures[3][getKeyFutures[0].size()+i].get());
}
for(int i = 0; i < getKeyFutures[2].size(); i++) {
ASSERT(getKeyFutures[2][i].get() == getKeyFutures[4][i].get());
}
for(int i = 0; i < getRangeFutures[0].size(); i++) {
if(getRangeFutures[0][i].get().size() != getRangeFutures[3][i].get().size()) {
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent(SevError, "SRL_ResultMismatch").detail("Size1", getRangeFutures[0][i].get().size()).detail("Size2", getRangeFutures[3][i].get().size());
Initial repository commit
2017-05-26 04:48:44 +08:00
for(auto kv : getRangeFutures[0][i].get())
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent("SRL_Result1").detail("Kv", printable(kv));
Initial repository commit
2017-05-26 04:48:44 +08:00
for(auto kv : getRangeFutures[3][i].get())
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent("SRL_Result2").detail("Kv", printable(kv));
Initial repository commit
2017-05-26 04:48:44 +08:00
ASSERT(false);
}
for(int j = 0; j < getRangeFutures[0][i].get().size(); j++) {
if(getRangeFutures[0][i].get()[j] != getRangeFutures[3][i].get()[j]) {
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent(SevError, "SRL_ResultMismatch").detail("J", j).detail("Result1", printable(getRangeFutures[0][i].get()[j])).detail("Result2", printable( getRangeFutures[3][i].get()[j]))
.detail("Result1Value", printable(getRangeFutures[0][i].get()[j].value)).detail("Result2Value", printable( getRangeFutures[3][i].get()[j].value));
Initial repository commit
2017-05-26 04:48:44 +08:00
for(auto kv : getRangeFutures[0][i].get())
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent("SRL_Result1").detail("Kv", printable(kv));
Initial repository commit
2017-05-26 04:48:44 +08:00
for(auto kv : getRangeFutures[3][i].get())
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent("SRL_Result2").detail("Kv", printable(kv));
Initial repository commit
2017-05-26 04:48:44 +08:00
ASSERT(false);
}
}
ASSERT(getRangeFutures[0][i].get() == getRangeFutures[3][i].get());
}
for(int i = 0; i < getRangeFutures[1].size(); i++) {
ASSERT(getRangeFutures[1][i].get() == getRangeFutures[3][getRangeFutures[0].size()+i].get());
}
for(int i = 0; i < getRangeFutures[2].size(); i++) {
if(getRangeFutures[2][i].get().size() != getRangeFutures[4][i].get().size()) {
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent(SevError, "SRL_ResultMismatch").detail("Size1", getRangeFutures[2][i].get().size()).detail("Size2", getRangeFutures[4][i].get().size());
Initial repository commit
2017-05-26 04:48:44 +08:00
for(auto kv : getRangeFutures[2][i].get())
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent("SRL_Result1").detail("Kv", printable(kv));
Initial repository commit
2017-05-26 04:48:44 +08:00
for(auto kv : getRangeFutures[4][i].get())
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent("SRL_Result2").detail("Kv", printable(kv));
Initial repository commit
2017-05-26 04:48:44 +08:00
ASSERT(false);
}
for(int j = 0; j < getRangeFutures[2][i].get().size(); j++) {
if(getRangeFutures[2][i].get()[j] != getRangeFutures[4][i].get()[j]) {
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent(SevError, "SRL_ResultMismatch").detail("J", j).detail("Result1", printable(getRangeFutures[2][i].get()[j])).detail("Result2", printable( getRangeFutures[4][i].get()[j]))
.detail("Result1Value", printable(getRangeFutures[2][i].get()[j].value)).detail("Result2Value", printable( getRangeFutures[4][i].get()[j].value));
Initial repository commit
2017-05-26 04:48:44 +08:00
for(auto kv : getRangeFutures[2][i].get())
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent("SRL_Result1").detail("Kv", printable(kv));
Initial repository commit
2017-05-26 04:48:44 +08:00
for(auto kv : getRangeFutures[4][i].get())
Attempt to normalize trace events: * Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check. * Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase. * Use seconds instead of milliseconds in details. Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed. This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
2018-06-09 02:11:08 +08:00
TraceEvent("SRL_Result2").detail("Kv", printable(kv));
Initial repository commit
2017-05-26 04:48:44 +08:00
ASSERT(false);
}
}
ASSERT(getRangeFutures[2][i].get() == getRangeFutures[4][i].get());
}
} catch( Error &e ) {
state ReadYourWritesTransaction trErr(cx);
Rewrite all `Void _ = wait(...)` -> `wait(...)`. This takes advantage of the new actorcompiler functionality to avoid having duplicate definitions of `Void _` when trying to feed the un-actorompiled source through clang.
2018-08-11 04:57:10 +08:00
wait( trErr.onError(e) );
Initial repository commit
2017-05-26 04:48:44 +08:00
}
}
}
};
WorkloadFactory<SerializabilityWorkload> SerializabilityWorkloadFactory("Serializability");