foundationdb/fdbserver/workloads/TenantManagementWorkload.ac...

1787 lines
66 KiB
C++

/*
* TenantManagementWorkload.actor.cpp
*
* This source file is part of the FoundationDB open source project
*
* Copyright 2013-2022 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.
*/
#include <cstdint>
#include <limits>
#include "fdbclient/ClientBooleanParams.h"
#include "fdbclient/ClusterConnectionMemoryRecord.h"
#include "fdbclient/FDBOptions.g.h"
#include "fdbclient/GenericManagementAPI.actor.h"
#include "fdbclient/KeyBackedTypes.h"
#include "fdbclient/MetaclusterManagement.actor.h"
#include "fdbclient/ReadYourWrites.h"
#include "fdbclient/RunTransaction.actor.h"
#include "fdbclient/TenantManagement.actor.h"
#include "fdbclient/TenantSpecialKeys.actor.h"
#include "fdbclient/ThreadSafeTransaction.h"
#include "fdbrpc/simulator.h"
#include "fdbserver/workloads/MetaclusterConsistency.actor.h"
#include "fdbserver/workloads/TenantConsistency.actor.h"
#include "fdbserver/workloads/workloads.actor.h"
#include "fdbserver/Knobs.h"
#include "flow/ApiVersion.h"
#include "flow/Error.h"
#include "flow/IRandom.h"
#include "flow/ThreadHelper.actor.h"
#include "flow/flow.h"
#include "libb64/decode.h"
#include "flow/actorcompiler.h" // This must be the last #include.
struct TenantManagementWorkload : TestWorkload {
struct TenantData {
int64_t id;
Optional<TenantGroupName> tenantGroup;
bool empty;
bool encrypted;
TenantData() : id(-1), empty(true) {}
TenantData(int64_t id, Optional<TenantGroupName> tenantGroup, bool empty, bool encrypted)
: id(id), tenantGroup(tenantGroup), empty(empty), encrypted(encrypted) {}
};
struct TenantGroupData {
int64_t tenantCount = 0;
};
std::map<TenantName, TenantData> createdTenants;
std::map<TenantGroupName, TenantGroupData> createdTenantGroups;
int64_t maxId = -1;
const Key keyName = "key"_sr;
const Key testParametersKey = "test_parameters"_sr;
const Value noTenantValue = "no_tenant"_sr;
const TenantName tenantNamePrefix = "tenant_management_workload_"_sr;
const ClusterName dataClusterName = "cluster1"_sr;
TenantName localTenantNamePrefix;
TenantName localTenantGroupNamePrefix;
const Key specialKeysTenantMapPrefix = SpecialKeySpace::getModuleRange(SpecialKeySpace::MODULE::MANAGEMENT)
.begin.withSuffix(TenantRangeImpl::submoduleRange.begin)
.withSuffix(TenantRangeImpl::mapSubRange.begin);
const Key specialKeysTenantConfigPrefix = SpecialKeySpace::getModuleRange(SpecialKeySpace::MODULE::MANAGEMENT)
.begin.withSuffix(TenantRangeImpl::submoduleRange.begin)
.withSuffix(TenantRangeImpl::configureSubRange.begin);
const Key specialKeysTenantRenamePrefix = SpecialKeySpace::getModuleRange(SpecialKeySpace::MODULE::MANAGEMENT)
.begin.withSuffix(TenantRangeImpl::submoduleRange.begin)
.withSuffix(TenantRangeImpl::renameSubRange.begin);
int maxTenants;
int maxTenantGroups;
double testDuration;
bool useMetacluster;
bool singleClient;
Version oldestDeletionVersion = 0;
Version newestDeletionVersion = 0;
Reference<IDatabase> mvDb;
Database dataDb;
// This test exercises multiple different ways to work with tenants
enum class OperationType {
// Use the special key-space APIs
SPECIAL_KEYS,
// Use the ManagementAPI functions that take a Database object and implement a retry loop
MANAGEMENT_DATABASE,
// Use the ManagementAPI functions that take a Transaction object
MANAGEMENT_TRANSACTION,
// Use the Metacluster API, if applicable. Note: not all APIs have a metacluster variant,
// and if there isn't one this will choose one of the other options.
METACLUSTER
};
OperationType randomOperationType() {
double metaclusterProb = useMetacluster ? 0.9 : 0.1;
if (deterministicRandom()->random01() < metaclusterProb) {
return OperationType::METACLUSTER;
} else {
return (OperationType)deterministicRandom()->randomInt(0, 3);
}
}
TenantManagementWorkload(WorkloadContext const& wcx) : TestWorkload(wcx) {
maxTenants = std::min<int>(1e8 - 1, getOption(options, "maxTenants"_sr, 1000));
maxTenantGroups = std::min<int>(2 * maxTenants, getOption(options, "maxTenantGroups"_sr, 20));
testDuration = getOption(options, "testDuration"_sr, 120.0);
singleClient = getOption(options, "singleClient"_sr, false);
localTenantNamePrefix = format("%stenant_%d_", tenantNamePrefix.toString().c_str(), clientId);
localTenantGroupNamePrefix = format("%stenantgroup_%d_", tenantNamePrefix.toString().c_str(), clientId);
bool defaultUseMetacluster = false;
if (clientId == 0 && g_network->isSimulated() && !g_simulator->extraDatabases.empty()) {
defaultUseMetacluster = deterministicRandom()->coinflip();
}
useMetacluster = getOption(options, "useMetacluster"_sr, defaultUseMetacluster);
}
std::string description() const override { return "TenantManagement"; }
struct TestParameters {
constexpr static FileIdentifier file_identifier = 1527576;
bool useMetacluster = false;
TestParameters() {}
TestParameters(bool useMetacluster) : useMetacluster(useMetacluster) {}
template <class Ar>
void serialize(Ar& ar) {
serializer(ar, useMetacluster);
}
Value encode() const { return ObjectWriter::toValue(*this, Unversioned()); }
static TestParameters decode(ValueRef const& value) {
return ObjectReader::fromStringRef<TestParameters>(value, Unversioned());
}
};
Future<Void> setup(Database const& cx) override {
if (clientId == 0 && g_network->isSimulated() && BUGGIFY) {
IKnobCollection::getMutableGlobalKnobCollection().setKnob(
"max_tenants_per_cluster", KnobValueRef::create(int{ deterministicRandom()->randomInt(20, 100) }));
}
if (clientId == 0 || !singleClient) {
return _setup(cx, this);
} else {
return Void();
}
}
ACTOR Future<Void> _setup(Database cx, TenantManagementWorkload* self) {
Reference<IDatabase> threadSafeHandle =
wait(unsafeThreadFutureToFuture(ThreadSafeDatabase::createFromExistingDatabase(cx)));
MultiVersionApi::api->selectApiVersion(cx->apiVersion.version());
self->mvDb = MultiVersionDatabase::debugCreateFromExistingDatabase(threadSafeHandle);
if (self->useMetacluster && self->clientId == 0) {
wait(success(MetaclusterAPI::createMetacluster(cx.getReference(), "management_cluster"_sr)));
DataClusterEntry entry;
entry.capacity.numTenantGroups = 1e9;
wait(MetaclusterAPI::registerCluster(
self->mvDb, self->dataClusterName, g_simulator->extraDatabases[0], entry));
}
state Transaction tr(cx);
if (self->clientId == 0) {
// Communicates test parameters to all other clients by storing it in a key
loop {
try {
tr.setOption(FDBTransactionOptions::RAW_ACCESS);
tr.set(self->testParametersKey, TestParameters(self->useMetacluster).encode());
wait(tr.commit());
break;
} catch (Error& e) {
wait(tr.onError(e));
}
}
} else {
// Read the parameters chosen and saved by client 0
loop {
try {
tr.setOption(FDBTransactionOptions::RAW_ACCESS);
Optional<Value> val = wait(tr.get(self->testParametersKey));
if (val.present()) {
TestParameters params = TestParameters::decode(val.get());
self->useMetacluster = params.useMetacluster;
break;
}
wait(delay(1.0));
tr.reset();
} catch (Error& e) {
wait(tr.onError(e));
}
}
}
if (self->useMetacluster) {
ASSERT(g_simulator->extraDatabases.size() == 1);
self->dataDb = Database::createSimulatedExtraDatabase(g_simulator->extraDatabases[0], cx->defaultTenant);
} else {
self->dataDb = cx;
}
if (self->clientId == 0) {
// Set a key outside of all tenants to make sure that our tenants aren't writing to the regular key-space
state Transaction dataTr(self->dataDb);
loop {
try {
dataTr.setOption(FDBTransactionOptions::RAW_ACCESS);
dataTr.set(self->keyName, self->noTenantValue);
wait(dataTr.commit());
break;
} catch (Error& e) {
wait(dataTr.onError(e));
}
}
}
return Void();
}
TenantName chooseTenantName(bool allowSystemTenant) {
TenantName tenant(format(
"%s%08d", localTenantNamePrefix.toString().c_str(), deterministicRandom()->randomInt(0, maxTenants)));
if (allowSystemTenant && deterministicRandom()->random01() < 0.02) {
tenant = tenant.withPrefix("\xff"_sr);
}
return tenant;
}
Optional<TenantGroupName> chooseTenantGroup(bool allowSystemTenantGroup, bool allowEmptyGroup = true) {
Optional<TenantGroupName> tenantGroup;
if (!allowEmptyGroup || deterministicRandom()->coinflip()) {
tenantGroup = TenantGroupNameRef(format("%s%08d",
localTenantGroupNamePrefix.toString().c_str(),
deterministicRandom()->randomInt(0, maxTenantGroups)));
if (allowSystemTenantGroup && deterministicRandom()->random01() < 0.02) {
tenantGroup = tenantGroup.get().withPrefix("\xff"_sr);
}
}
return tenantGroup;
}
Future<Optional<TenantMapEntry>> tryGetTenant(TenantName tenantName, OperationType operationType) {
if (operationType == OperationType::METACLUSTER) {
return MetaclusterAPI::tryGetTenant(mvDb, tenantName);
} else {
return TenantAPI::tryGetTenant(dataDb.getReference(), tenantName);
}
}
// Creates tenant(s) using the specified operation type
ACTOR static Future<Void> createTenantImpl(Reference<ReadYourWritesTransaction> tr,
std::map<TenantName, TenantMapEntry> tenantsToCreate,
OperationType operationType,
TenantManagementWorkload* self) {
if (operationType == OperationType::SPECIAL_KEYS) {
tr->setOption(FDBTransactionOptions::SPECIAL_KEY_SPACE_ENABLE_WRITES);
for (auto [tenant, entry] : tenantsToCreate) {
tr->set(self->specialKeysTenantMapPrefix.withSuffix(tenant), ""_sr);
if (entry.tenantGroup.present()) {
tr->set(self->specialKeysTenantConfigPrefix.withSuffix(
Tuple::makeTuple(tenant, "tenant_group"_sr).pack()),
entry.tenantGroup.get());
}
}
wait(tr->commit());
} else if (operationType == OperationType::MANAGEMENT_DATABASE) {
ASSERT(tenantsToCreate.size() == 1);
wait(success(TenantAPI::createTenant(
self->dataDb.getReference(), tenantsToCreate.begin()->first, tenantsToCreate.begin()->second)));
} else if (operationType == OperationType::MANAGEMENT_TRANSACTION) {
tr->setOption(FDBTransactionOptions::ACCESS_SYSTEM_KEYS);
int64_t _nextId = wait(TenantAPI::getNextTenantId(tr));
int64_t nextId = _nextId;
std::vector<Future<Void>> createFutures;
for (auto [tenant, entry] : tenantsToCreate) {
entry.setId(nextId++);
createFutures.push_back(success(TenantAPI::createTenantTransaction(tr, tenant, entry)));
}
TenantMetadata::lastTenantId().set(tr, nextId - 1);
wait(waitForAll(createFutures));
wait(tr->commit());
} else {
ASSERT(tenantsToCreate.size() == 1);
if (deterministicRandom()->coinflip()) {
tenantsToCreate.begin()->second.assignedCluster = self->dataClusterName;
}
wait(MetaclusterAPI::createTenant(
self->mvDb, tenantsToCreate.begin()->first, tenantsToCreate.begin()->second));
}
return Void();
}
ACTOR static Future<Void> createTenant(TenantManagementWorkload* self) {
state OperationType operationType = self->randomOperationType();
int numTenants = 1;
// For transaction-based operations, test creating multiple tenants in the same transaction
if (operationType == OperationType::SPECIAL_KEYS || operationType == OperationType::MANAGEMENT_TRANSACTION) {
numTenants = deterministicRandom()->randomInt(1, 5);
}
// Tracks whether any tenant exists in the database or not. This variable is updated if we have to retry
// the creation.
state bool alreadyExists = false;
// True if any tenant name starts with \xff
state bool hasSystemTenant = false;
// True if any tenant group name starts with \xff
state bool hasSystemTenantGroup = false;
state int newTenants = 0;
state std::map<TenantName, TenantMapEntry> tenantsToCreate;
for (int i = 0; i < numTenants; ++i) {
TenantName tenant = self->chooseTenantName(true);
while (tenantsToCreate.count(tenant)) {
tenant = self->chooseTenantName(true);
}
TenantMapEntry entry;
entry.tenantGroup = self->chooseTenantGroup(true);
if (operationType == OperationType::SPECIAL_KEYS) {
entry.encrypted = SERVER_KNOBS->ENABLE_ENCRYPTION;
} else {
entry.encrypted = deterministicRandom()->coinflip();
}
if (self->createdTenants.count(tenant)) {
alreadyExists = true;
} else if (!tenantsToCreate.count(tenant)) {
++newTenants;
}
tenantsToCreate[tenant] = entry;
hasSystemTenant = hasSystemTenant || tenant.startsWith("\xff"_sr);
hasSystemTenantGroup = hasSystemTenantGroup || entry.tenantGroup.orDefault(""_sr).startsWith("\xff"_sr);
}
// If any tenant existed at the start of this function, then we expect the creation to fail or be a no-op,
// depending on the type of create operation being executed
state bool existedAtStart = alreadyExists;
state Reference<ReadYourWritesTransaction> tr = makeReference<ReadYourWritesTransaction>(self->dataDb);
state int64_t minTenantCount = std::numeric_limits<int64_t>::max();
state int64_t finalTenantCount = 0;
loop {
try {
// First, attempt to create the tenants
state bool retried = false;
loop {
if (operationType == OperationType::MANAGEMENT_TRANSACTION ||
operationType == OperationType::SPECIAL_KEYS) {
tr->setOption(FDBTransactionOptions::READ_SYSTEM_KEYS);
wait(store(finalTenantCount, TenantMetadata::tenantCount().getD(tr, Snapshot::False, 0)));
minTenantCount = std::min(finalTenantCount, minTenantCount);
}
try {
Optional<Void> result = wait(timeout(createTenantImpl(tr, tenantsToCreate, operationType, self),
deterministicRandom()->randomInt(1, 30)));
if (result.present()) {
// Make sure that we had capacity to create the tenants. This cannot be validated for
// database operations because we cannot determine the tenant count in the same transaction
// that the tenant is created
if (operationType == OperationType::SPECIAL_KEYS ||
operationType == OperationType::MANAGEMENT_TRANSACTION) {
ASSERT(minTenantCount + newTenants <= CLIENT_KNOBS->MAX_TENANTS_PER_CLUSTER);
} else {
// Database operations shouldn't get here if the tenant already exists
ASSERT(!alreadyExists);
}
break;
}
retried = true;
tr->reset();
} catch (Error& e) {
// If we retried the creation after our initial attempt succeeded, then we proceed with the rest
// of the creation steps normally. Otherwise, the creation happened elsewhere and we failed
// here, so we can rethrow the error.
if (e.code() == error_code_tenant_already_exists && !existedAtStart) {
ASSERT(operationType == OperationType::METACLUSTER ||
operationType == OperationType::MANAGEMENT_DATABASE);
ASSERT(retried);
break;
} else {
throw;
}
}
// Check the state of the first created tenant
Optional<TenantMapEntry> resultEntry =
wait(self->tryGetTenant(tenantsToCreate.begin()->first, operationType));
if (resultEntry.present()) {
if (resultEntry.get().tenantState == TenantState::READY) {
// The tenant now exists, so we will retry and expect the creation to react accordingly
alreadyExists = true;
} else {
// Only a metacluster tenant creation can end up in a partially created state
// We should be able to retry and pick up where we left off
ASSERT(operationType == OperationType::METACLUSTER);
ASSERT(resultEntry.get().tenantState == TenantState::REGISTERING);
}
}
}
// Check that using the wrong creation type fails depending on whether we are using a metacluster
ASSERT(self->useMetacluster == (operationType == OperationType::METACLUSTER));
// Database-based creation modes will fail if the tenant already existed
if (operationType == OperationType::MANAGEMENT_DATABASE ||
operationType == OperationType::METACLUSTER) {
ASSERT(!existedAtStart);
}
// It is not legal to create a tenant or tenant group starting with \xff
ASSERT(!hasSystemTenant);
ASSERT(!hasSystemTenantGroup);
state std::map<TenantName, TenantMapEntry>::iterator tenantItr;
for (tenantItr = tenantsToCreate.begin(); tenantItr != tenantsToCreate.end(); ++tenantItr) {
// Ignore any tenants that already existed
if (self->createdTenants.count(tenantItr->first)) {
continue;
}
// Read the created tenant object and verify that its state is correct
state Optional<TenantMapEntry> entry = wait(self->tryGetTenant(tenantItr->first, operationType));
ASSERT(entry.present());
ASSERT(entry.get().id > self->maxId);
ASSERT(entry.get().tenantGroup == tenantItr->second.tenantGroup);
ASSERT(entry.get().tenantState == TenantState::READY);
if (self->useMetacluster) {
// In a metacluster, we should also see that the tenant was created on the data cluster
Optional<TenantMapEntry> dataEntry =
wait(TenantAPI::tryGetTenant(self->dataDb.getReference(), tenantItr->first));
ASSERT(dataEntry.present());
ASSERT(dataEntry.get().id == entry.get().id);
ASSERT(dataEntry.get().tenantGroup == entry.get().tenantGroup);
ASSERT(dataEntry.get().tenantState == TenantState::READY);
}
// Update our local tenant state to include the newly created one
self->maxId = entry.get().id;
self->createdTenants[tenantItr->first] =
TenantData(entry.get().id, tenantItr->second.tenantGroup, true, tenantItr->second.encrypted);
// If this tenant has a tenant group, create or update the entry for it
if (tenantItr->second.tenantGroup.present()) {
self->createdTenantGroups[tenantItr->second.tenantGroup.get()].tenantCount++;
}
// Randomly decide to insert a key into the tenant
state bool insertData = deterministicRandom()->random01() < 0.5;
if (insertData) {
state Transaction insertTr(self->dataDb, tenantItr->first);
loop {
try {
// The value stored in the key will be the name of the tenant
insertTr.set(self->keyName, tenantItr->first);
wait(insertTr.commit());
break;
} catch (Error& e) {
wait(insertTr.onError(e));
}
}
self->createdTenants[tenantItr->first].empty = false;
// Make sure that the key inserted correctly concatenates the tenant prefix with the
// relative key
state Transaction checkTr(self->dataDb);
loop {
try {
checkTr.setOption(FDBTransactionOptions::RAW_ACCESS);
Optional<Value> val = wait(checkTr.get(self->keyName.withPrefix(entry.get().prefix)));
ASSERT(val.present());
ASSERT(val.get() == tenantItr->first);
break;
} catch (Error& e) {
wait(checkTr.onError(e));
}
}
}
// Perform some final tenant validation
wait(checkTenantContents(self, tenantItr->first, self->createdTenants[tenantItr->first]));
}
return Void();
} catch (Error& e) {
if (e.code() == error_code_invalid_tenant_name) {
ASSERT(hasSystemTenant);
return Void();
} else if (e.code() == error_code_invalid_tenant_group_name) {
ASSERT(hasSystemTenantGroup);
return Void();
} else if (e.code() == error_code_invalid_metacluster_operation) {
ASSERT(operationType == OperationType::METACLUSTER != self->useMetacluster);
return Void();
} else if (e.code() == error_code_cluster_no_capacity) {
// Confirm that we overshot our capacity. This check cannot be done for database operations
// because we cannot transactionally get the tenant count with the creation.
if (operationType == OperationType::MANAGEMENT_TRANSACTION ||
operationType == OperationType::SPECIAL_KEYS) {
ASSERT(finalTenantCount + newTenants > CLIENT_KNOBS->MAX_TENANTS_PER_CLUSTER);
}
return Void();
}
// Database-based operations should not need to be retried
else if (operationType == OperationType::MANAGEMENT_DATABASE ||
operationType == OperationType::METACLUSTER) {
if (e.code() == error_code_tenant_already_exists) {
ASSERT(existedAtStart);
} else {
ASSERT(tenantsToCreate.size() == 1);
TraceEvent(SevError, "CreateTenantFailure")
.error(e)
.detail("TenantName", tenantsToCreate.begin()->first);
}
return Void();
}
// Transaction-based operations should be retried
else {
try {
wait(tr->onError(e));
} catch (Error& e) {
for (auto [tenant, _] : tenantsToCreate) {
TraceEvent(SevError, "CreateTenantFailure").error(e).detail("TenantName", tenant);
}
return Void();
}
}
}
}
}
// Deletes the tenant or tenant range using the specified operation type
ACTOR static Future<Void> deleteTenantImpl(Reference<ReadYourWritesTransaction> tr,
TenantName beginTenant,
Optional<TenantName> endTenant,
std::vector<TenantName> tenants,
OperationType operationType,
TenantManagementWorkload* self) {
state int tenantIndex;
if (operationType == OperationType::SPECIAL_KEYS) {
tr->setOption(FDBTransactionOptions::SPECIAL_KEY_SPACE_ENABLE_WRITES);
Key key = self->specialKeysTenantMapPrefix.withSuffix(beginTenant);
if (endTenant.present()) {
tr->clear(KeyRangeRef(key, self->specialKeysTenantMapPrefix.withSuffix(endTenant.get())));
} else {
tr->clear(key);
}
wait(tr->commit());
} else if (operationType == OperationType::MANAGEMENT_DATABASE) {
ASSERT(!endTenant.present() && tenants.size() == 1);
wait(TenantAPI::deleteTenant(self->dataDb.getReference(), beginTenant));
} else if (operationType == OperationType::MANAGEMENT_TRANSACTION) {
tr->setOption(FDBTransactionOptions::ACCESS_SYSTEM_KEYS);
std::vector<Future<Void>> deleteFutures;
for (tenantIndex = 0; tenantIndex != tenants.size(); ++tenantIndex) {
deleteFutures.push_back(TenantAPI::deleteTenantTransaction(tr, tenants[tenantIndex]));
}
wait(waitForAll(deleteFutures));
wait(tr->commit());
} else {
ASSERT(!endTenant.present() && tenants.size() == 1);
wait(MetaclusterAPI::deleteTenant(self->mvDb, beginTenant));
}
return Void();
}
// Returns GRV and eats GRV errors
ACTOR static Future<Version> getReadVersion(Reference<ReadYourWritesTransaction> tr) {
loop {
try {
Version version = wait(tr->getReadVersion());
return version;
} catch (Error& e) {
if (e.code() == error_code_grv_proxy_memory_limit_exceeded ||
e.code() == error_code_batch_transaction_throttled) {
wait(tr->onError(e));
} else {
throw;
}
}
}
}
ACTOR static Future<Void> deleteTenant(TenantManagementWorkload* self) {
state TenantName beginTenant = self->chooseTenantName(true);
state OperationType operationType = self->randomOperationType();
state Reference<ReadYourWritesTransaction> tr = makeReference<ReadYourWritesTransaction>(self->dataDb);
// For transaction-based deletion, we randomly allow the deletion of a range of tenants
state Optional<TenantName> endTenant =
operationType != OperationType::MANAGEMENT_DATABASE && operationType != OperationType::METACLUSTER &&
!beginTenant.startsWith("\xff"_sr) && deterministicRandom()->random01() < 0.2
? Optional<TenantName>(self->chooseTenantName(false))
: Optional<TenantName>();
if (endTenant.present() && endTenant < beginTenant) {
TenantName temp = beginTenant;
beginTenant = endTenant.get();
endTenant = temp;
}
auto itr = self->createdTenants.find(beginTenant);
// True if the beginTenant should exist and be deletable. This is updated if a deletion fails and gets
// retried.
state bool alreadyExists = itr != self->createdTenants.end();
// True if the beginTenant existed at the start of this function
state bool existedAtStart = alreadyExists;
// True if all of the tenants in the range are empty and can be deleted
state bool isEmpty = true;
// True if we expect that some tenant will be deleted
state bool anyExists = alreadyExists;
// Collect a list of all tenants that we expect should be deleted by this operation
state std::vector<TenantName> tenants;
if (!endTenant.present()) {
tenants.push_back(beginTenant);
} else if (endTenant.present()) {
for (auto itr = self->createdTenants.lower_bound(beginTenant);
itr != self->createdTenants.end() && itr->first < endTenant.get();
++itr) {
tenants.push_back(itr->first);
anyExists = true;
}
}
// Check whether each tenant is empty.
state int tenantIndex;
try {
if (alreadyExists || endTenant.present()) {
for (tenantIndex = 0; tenantIndex < tenants.size(); ++tenantIndex) {
// For most tenants, we will delete the contents and make them empty
if (deterministicRandom()->random01() < 0.9) {
state Transaction clearTr(self->dataDb, tenants[tenantIndex]);
loop {
try {
clearTr.clear(self->keyName);
wait(clearTr.commit());
auto itr = self->createdTenants.find(tenants[tenantIndex]);
ASSERT(itr != self->createdTenants.end());
itr->second.empty = true;
break;
} catch (Error& e) {
wait(clearTr.onError(e));
}
}
}
// Otherwise, we will just report the current emptiness of the tenant
else {
auto itr = self->createdTenants.find(tenants[tenantIndex]);
ASSERT(itr != self->createdTenants.end());
isEmpty = isEmpty && itr->second.empty;
}
}
}
} catch (Error& e) {
TraceEvent(SevError, "DeleteTenantFailure")
.error(e)
.detail("TenantName", beginTenant)
.detail("EndTenant", endTenant);
return Void();
}
loop {
try {
// Attempt to delete the tenant(s)
state bool retried = false;
loop {
try {
state Version beforeVersion = wait(self->getReadVersion(tr));
Optional<Void> result =
wait(timeout(deleteTenantImpl(tr, beginTenant, endTenant, tenants, operationType, self),
deterministicRandom()->randomInt(1, 30)));
if (result.present()) {
if (anyExists) {
if (self->oldestDeletionVersion == 0 && !tenants.empty()) {
tr->reset();
Version afterVersion = wait(self->getReadVersion(tr));
self->oldestDeletionVersion = afterVersion;
}
self->newestDeletionVersion = beforeVersion;
}
// Database operations shouldn't get here if the tenant didn't exist
ASSERT(operationType == OperationType::SPECIAL_KEYS ||
operationType == OperationType::MANAGEMENT_TRANSACTION || alreadyExists);
break;
}
retried = true;
tr->reset();
} catch (Error& e) {
// If we retried the deletion after our initial attempt succeeded, then we proceed with the
// rest of the deletion steps normally. Otherwise, the deletion happened elsewhere and we
// failed here, so we can rethrow the error.
if (e.code() == error_code_tenant_not_found && existedAtStart) {
ASSERT(operationType == OperationType::METACLUSTER ||
operationType == OperationType::MANAGEMENT_DATABASE);
ASSERT(retried);
break;
} else if (e.code() == error_code_grv_proxy_memory_limit_exceeded ||
e.code() == error_code_batch_transaction_throttled) {
// GRV proxy returns an error
wait(tr->onError(e));
continue;
} else {
throw;
}
}
if (!tenants.empty()) {
// Check the state of the first deleted tenant
Optional<TenantMapEntry> resultEntry =
wait(self->tryGetTenant(*tenants.begin(), operationType));
if (!resultEntry.present()) {
alreadyExists = false;
} else if (resultEntry.get().tenantState == TenantState::REMOVING) {
ASSERT(operationType == OperationType::METACLUSTER);
} else {
ASSERT(resultEntry.get().tenantState == TenantState::READY);
}
}
}
// The management transaction operation is a no-op if there are no tenants to delete in a range
// delete
if (tenants.size() == 0 && operationType == OperationType::MANAGEMENT_TRANSACTION) {
return Void();
}
// The special keys operation is a no-op if the begin and end tenant are equal (i.e. the range is
// empty)
if (endTenant.present() && beginTenant == endTenant.get() &&
operationType == OperationType::SPECIAL_KEYS) {
return Void();
}
// Check that using the wrong deletion type fails depending on whether we are using a metacluster
ASSERT(self->useMetacluster == (operationType == OperationType::METACLUSTER));
// Transaction-based operations do not fail if the tenant isn't present. If we attempted to delete a
// single tenant that didn't exist, we can just return.
if (!existedAtStart && !endTenant.present() &&
(operationType == OperationType::MANAGEMENT_TRANSACTION ||
operationType == OperationType::SPECIAL_KEYS)) {
return Void();
}
ASSERT(existedAtStart || endTenant.present());
// Deletion should not succeed if any tenant in the range wasn't empty
ASSERT(isEmpty);
// Update our local state to remove the deleted tenants
for (auto tenant : tenants) {
auto itr = self->createdTenants.find(tenant);
ASSERT(itr != self->createdTenants.end());
// If the tenant group has no tenants remaining, stop tracking it
if (itr->second.tenantGroup.present()) {
auto tenantGroupItr = self->createdTenantGroups.find(itr->second.tenantGroup.get());
ASSERT(tenantGroupItr != self->createdTenantGroups.end());
if (--tenantGroupItr->second.tenantCount == 0) {
self->createdTenantGroups.erase(tenantGroupItr);
}
}
self->createdTenants.erase(tenant);
}
return Void();
} catch (Error& e) {
if (e.code() == error_code_tenant_not_empty) {
ASSERT(!isEmpty);
return Void();
} else if (e.code() == error_code_invalid_metacluster_operation) {
ASSERT(operationType == OperationType::METACLUSTER != self->useMetacluster);
return Void();
}
// Database-based operations do not need to be retried
else if (operationType == OperationType::MANAGEMENT_DATABASE ||
operationType == OperationType::METACLUSTER) {
if (e.code() == error_code_tenant_not_found) {
ASSERT(!existedAtStart && !endTenant.present());
} else {
TraceEvent(SevError, "DeleteTenantFailure")
.error(e)
.detail("TenantName", beginTenant)
.detail("EndTenant", endTenant);
}
return Void();
}
// Transaction-based operations should be retried
else {
try {
wait(tr->onError(e));
} catch (Error& e) {
TraceEvent(SevError, "DeleteTenantFailure")
.error(e)
.detail("TenantName", beginTenant)
.detail("EndTenant", endTenant);
return Void();
}
}
}
}
}
// Performs some validation on a tenant's contents
ACTOR static Future<Void> checkTenantContents(TenantManagementWorkload* self,
TenantName tenant,
TenantData tenantData) {
state Transaction tr(self->dataDb, tenant);
loop {
try {
// We only every store a single key in each tenant. Therefore we expect a range read of the entire
// tenant to return either 0 or 1 keys, depending on whether that key has been set.
state RangeResult result = wait(tr.getRange(KeyRangeRef(""_sr, "\xff"_sr), 2));
// An empty tenant should have no data
if (tenantData.empty) {
ASSERT(result.size() == 0);
}
// A non-empty tenant should have our single key. The value of that key should be the name of the
// tenant.
else {
ASSERT(result.size() == 1);
ASSERT(result[0].key == self->keyName);
ASSERT(result[0].value == tenant);
}
break;
} catch (Error& e) {
wait(tr.onError(e));
}
}
return Void();
}
// Convert the JSON document returned by the special-key space when reading tenant metadata
// into a TenantMapEntry
static TenantMapEntry jsonToTenantMapEntry(ValueRef tenantJson) {
json_spirit::mValue jsonObject;
json_spirit::read_string(tenantJson.toString(), jsonObject);
JSONDoc jsonDoc(jsonObject);
int64_t id;
std::string prefix;
std::string base64Prefix;
std::string printablePrefix;
std::string tenantStateStr;
std::string base64TenantGroup;
std::string printableTenantGroup;
bool encrypted;
std::string assignedClusterStr;
jsonDoc.get("id", id);
jsonDoc.get("prefix.base64", base64Prefix);
jsonDoc.get("prefix.printable", printablePrefix);
jsonDoc.get("prefix.encrypted", encrypted);
prefix = base64::decoder::from_string(base64Prefix);
ASSERT(prefix == unprintable(printablePrefix));
jsonDoc.get("tenant_state", tenantStateStr);
Optional<TenantGroupName> tenantGroup;
if (jsonDoc.tryGet("tenant_group.base64", base64TenantGroup)) {
jsonDoc.get("tenant_group.printable", printableTenantGroup);
std::string tenantGroupStr = base64::decoder::from_string(base64TenantGroup);
ASSERT(tenantGroupStr == unprintable(printableTenantGroup));
tenantGroup = TenantGroupNameRef(tenantGroupStr);
}
Optional<ClusterName> assignedCluster;
if (jsonDoc.tryGet("assigned_cluster", assignedClusterStr)) {
assignedCluster = ClusterNameRef(assignedClusterStr);
}
TenantMapEntry entry(id, TenantMapEntry::stringToTenantState(tenantStateStr), tenantGroup, encrypted);
ASSERT(entry.prefix == prefix);
return entry;
}
// Gets the metadata for a tenant using the specified operation type
ACTOR static Future<TenantMapEntry> getTenantImpl(Reference<ReadYourWritesTransaction> tr,
TenantName tenant,
OperationType operationType,
TenantManagementWorkload* self) {
state TenantMapEntry entry;
if (operationType == OperationType::SPECIAL_KEYS) {
Key key = self->specialKeysTenantMapPrefix.withSuffix(tenant);
Optional<Value> value = wait(tr->get(key));
if (!value.present()) {
throw tenant_not_found();
}
entry = TenantManagementWorkload::jsonToTenantMapEntry(value.get());
} else if (operationType == OperationType::MANAGEMENT_DATABASE) {
wait(store(entry, TenantAPI::getTenant(self->dataDb.getReference(), tenant)));
} else if (operationType == OperationType::MANAGEMENT_TRANSACTION) {
tr->setOption(FDBTransactionOptions::READ_SYSTEM_KEYS);
wait(store(entry, TenantAPI::getTenantTransaction(tr, tenant)));
} else {
wait(store(entry, MetaclusterAPI::getTenant(self->mvDb, tenant)));
}
return entry;
}
ACTOR static Future<Void> getTenant(TenantManagementWorkload* self) {
state TenantName tenant = self->chooseTenantName(true);
state OperationType operationType = self->randomOperationType();
state Reference<ReadYourWritesTransaction> tr = makeReference<ReadYourWritesTransaction>(self->dataDb);
// True if the tenant should should exist and return a result
auto itr = self->createdTenants.find(tenant);
state bool alreadyExists = itr != self->createdTenants.end() &&
!(operationType == OperationType::METACLUSTER && !self->useMetacluster);
state TenantData tenantData = alreadyExists ? itr->second : TenantData();
loop {
try {
// Get the tenant metadata and check that it matches our local state
state TenantMapEntry entry = wait(getTenantImpl(tr, tenant, operationType, self));
ASSERT(alreadyExists);
ASSERT(entry.id == tenantData.id);
ASSERT(entry.tenantGroup == tenantData.tenantGroup);
wait(checkTenantContents(self, tenant, tenantData));
return Void();
} catch (Error& e) {
state bool retry = false;
state Error error = e;
if (e.code() == error_code_tenant_not_found) {
ASSERT(!alreadyExists);
return Void();
}
// Transaction-based operations should retry
else if (operationType == OperationType::MANAGEMENT_TRANSACTION ||
operationType == OperationType::SPECIAL_KEYS) {
try {
retry = true;
wait(tr->onError(e));
retry = true;
} catch (Error& e) {
error = e;
retry = false;
}
}
if (!retry) {
TraceEvent(SevError, "GetTenantFailure").error(error).detail("TenantName", tenant);
return Void();
}
}
}
}
// Gets a list of tenants using the specified operation type
ACTOR static Future<std::vector<std::pair<TenantName, TenantMapEntry>>> listTenantsImpl(
Reference<ReadYourWritesTransaction> tr,
TenantName beginTenant,
TenantName endTenant,
int limit,
OperationType operationType,
TenantManagementWorkload* self) {
state std::vector<std::pair<TenantName, TenantMapEntry>> tenants;
if (operationType == OperationType::SPECIAL_KEYS) {
KeyRange range = KeyRangeRef(beginTenant, endTenant).withPrefix(self->specialKeysTenantMapPrefix);
RangeResult results = wait(tr->getRange(range, limit));
for (auto result : results) {
tenants.push_back(std::make_pair(result.key.removePrefix(self->specialKeysTenantMapPrefix),
TenantManagementWorkload::jsonToTenantMapEntry(result.value)));
}
} else if (operationType == OperationType::MANAGEMENT_DATABASE) {
wait(store(tenants, TenantAPI::listTenants(self->dataDb.getReference(), beginTenant, endTenant, limit)));
} else if (operationType == OperationType::MANAGEMENT_TRANSACTION) {
tr->setOption(FDBTransactionOptions::READ_SYSTEM_KEYS);
wait(store(tenants, TenantAPI::listTenantsTransaction(tr, beginTenant, endTenant, limit)));
} else {
wait(store(tenants, MetaclusterAPI::listTenants(self->mvDb, beginTenant, endTenant, limit)));
}
return tenants;
}
ACTOR static Future<Void> listTenants(TenantManagementWorkload* self) {
state TenantName beginTenant = self->chooseTenantName(false);
state TenantName endTenant = self->chooseTenantName(false);
state int limit = std::min(CLIENT_KNOBS->MAX_TENANTS_PER_CLUSTER + 1,
deterministicRandom()->randomInt(1, self->maxTenants * 2));
state OperationType operationType = self->randomOperationType();
state Reference<ReadYourWritesTransaction> tr = makeReference<ReadYourWritesTransaction>(self->dataDb);
if (beginTenant > endTenant) {
std::swap(beginTenant, endTenant);
}
loop {
try {
// Attempt to read the chosen list of tenants
state std::vector<std::pair<TenantName, TenantMapEntry>> tenants =
wait(listTenantsImpl(tr, beginTenant, endTenant, limit, operationType, self));
// Attempting to read the list of tenants using the metacluster API in a non-metacluster should
// return nothing in this test
if (operationType == OperationType::METACLUSTER && !self->useMetacluster) {
ASSERT(tenants.size() == 0);
return Void();
}
ASSERT(tenants.size() <= limit);
// Compare the resulting tenant list to the list we expected to get
auto localItr = self->createdTenants.lower_bound(beginTenant);
auto tenantMapItr = tenants.begin();
for (; tenantMapItr != tenants.end(); ++tenantMapItr, ++localItr) {
ASSERT(localItr != self->createdTenants.end());
ASSERT(localItr->first == tenantMapItr->first);
}
// Make sure the list terminated at the right spot
ASSERT(tenants.size() == limit || localItr == self->createdTenants.end() ||
localItr->first >= endTenant);
return Void();
} catch (Error& e) {
state bool retry = false;
state Error error = e;
// Transaction-based operations need to be retried
if (operationType == OperationType::MANAGEMENT_TRANSACTION ||
operationType == OperationType::SPECIAL_KEYS) {
try {
retry = true;
wait(tr->onError(e));
} catch (Error& e) {
error = e;
retry = false;
}
}
if (!retry) {
TraceEvent(SevError, "ListTenantFailure")
.error(error)
.detail("BeginTenant", beginTenant)
.detail("EndTenant", endTenant);
return Void();
}
}
}
}
// Helper function that checks tenant keyspace and updates internal Tenant Map after a rename operation
ACTOR static Future<Void> verifyTenantRename(TenantManagementWorkload* self,
TenantName oldTenantName,
TenantName newTenantName) {
state Optional<TenantMapEntry> oldTenantEntry =
wait(TenantAPI::tryGetTenant(self->dataDb.getReference(), oldTenantName));
state Optional<TenantMapEntry> newTenantEntry =
wait(TenantAPI::tryGetTenant(self->dataDb.getReference(), newTenantName));
ASSERT(!oldTenantEntry.present());
ASSERT(newTenantEntry.present());
TenantData tData = self->createdTenants[oldTenantName];
self->createdTenants[newTenantName] = tData;
self->createdTenants.erase(oldTenantName);
state Transaction insertTr(self->dataDb, newTenantName);
if (!tData.empty) {
loop {
try {
insertTr.set(self->keyName, newTenantName);
wait(insertTr.commit());
break;
} catch (Error& e) {
wait(insertTr.onError(e));
}
}
}
return Void();
}
ACTOR static Future<Void> verifyTenantRenames(TenantManagementWorkload* self,
std::map<TenantName, TenantName> tenantRenames) {
state std::map<TenantName, TenantName> tenantRenamesCopy = tenantRenames;
state std::map<TenantName, TenantName>::iterator iter = tenantRenamesCopy.begin();
for (; iter != tenantRenamesCopy.end(); ++iter) {
wait(verifyTenantRename(self, iter->first, iter->second));
wait(checkTenantContents(self, iter->second, self->createdTenants[iter->second]));
}
return Void();
}
ACTOR static Future<Void> renameTenantImpl(Reference<ReadYourWritesTransaction> tr,
OperationType operationType,
std::map<TenantName, TenantName> tenantRenames,
bool tenantNotFound,
bool tenantExists,
bool tenantOverlap,
TenantManagementWorkload* self) {
if (operationType == OperationType::SPECIAL_KEYS) {
tr->setOption(FDBTransactionOptions::SPECIAL_KEY_SPACE_ENABLE_WRITES);
for (auto& iter : tenantRenames) {
tr->set(self->specialKeysTenantRenamePrefix.withSuffix(iter.first), iter.second);
}
wait(tr->commit());
} else if (operationType == OperationType::MANAGEMENT_DATABASE) {
ASSERT(tenantRenames.size() == 1);
auto iter = tenantRenames.begin();
wait(TenantAPI::renameTenant(self->dataDb.getReference(), iter->first, iter->second));
ASSERT(!tenantNotFound && !tenantExists);
} else if (operationType == OperationType::MANAGEMENT_TRANSACTION) {
tr->setOption(FDBTransactionOptions::ACCESS_SYSTEM_KEYS);
std::vector<Future<Void>> renameFutures;
for (auto& iter : tenantRenames) {
renameFutures.push_back(success(TenantAPI::renameTenantTransaction(tr, iter.first, iter.second)));
}
wait(waitForAll(renameFutures));
wait(tr->commit());
ASSERT(!tenantNotFound && !tenantExists);
} else { // operationType == OperationType::METACLUSTER
ASSERT(tenantRenames.size() == 1);
auto iter = tenantRenames.begin();
wait(MetaclusterAPI::renameTenant(self->mvDb, iter->first, iter->second));
}
return Void();
}
ACTOR static Future<Void> renameTenant(TenantManagementWorkload* self) {
state OperationType operationType = self->randomOperationType();
state int numTenants = 1;
state Reference<ReadYourWritesTransaction> tr = self->dataDb->createTransaction();
if (operationType == OperationType::SPECIAL_KEYS || operationType == OperationType::MANAGEMENT_TRANSACTION) {
numTenants = deterministicRandom()->randomInt(1, 5);
}
state std::map<TenantName, TenantName> tenantRenames;
state std::set<TenantName> allTenantNames;
// Tenant Error flags
state bool tenantExists = false;
state bool tenantNotFound = false;
state bool tenantOverlap = false;
state bool unknownResult = false;
for (int i = 0; i < numTenants; ++i) {
TenantName oldTenant = self->chooseTenantName(false);
TenantName newTenant = self->chooseTenantName(false);
bool checkOverlap =
oldTenant == newTenant || allTenantNames.count(oldTenant) || allTenantNames.count(newTenant);
// renameTenantTransaction does not handle rename collisions:
// reject the rename here if it has overlap and we are doing a transaction operation
// and then pick another combination
if (checkOverlap && operationType == OperationType::MANAGEMENT_TRANSACTION) {
--i;
continue;
}
tenantOverlap = tenantOverlap || checkOverlap;
tenantRenames[oldTenant] = newTenant;
allTenantNames.insert(oldTenant);
allTenantNames.insert(newTenant);
if (!self->createdTenants.count(oldTenant)) {
tenantNotFound = true;
}
if (self->createdTenants.count(newTenant)) {
tenantExists = true;
}
}
loop {
try {
wait(renameTenantImpl(
tr, operationType, tenantRenames, tenantNotFound, tenantExists, tenantOverlap, self));
wait(verifyTenantRenames(self, tenantRenames));
// Check that using the wrong rename API fails depending on whether we are using a metacluster
ASSERT(self->useMetacluster == (operationType == OperationType::METACLUSTER));
return Void();
} catch (Error& e) {
if (e.code() == error_code_tenant_not_found) {
if (unknownResult) {
wait(verifyTenantRenames(self, tenantRenames));
} else {
ASSERT(tenantNotFound);
}
return Void();
} else if (e.code() == error_code_tenant_already_exists) {
if (unknownResult) {
wait(verifyTenantRenames(self, tenantRenames));
} else {
ASSERT(tenantExists);
}
return Void();
} else if (e.code() == error_code_special_keys_api_failure) {
ASSERT(tenantOverlap);
return Void();
} else if (e.code() == error_code_invalid_metacluster_operation) {
ASSERT(operationType == OperationType::METACLUSTER != self->useMetacluster);
return Void();
} else if (e.code() == error_code_cluster_no_capacity) {
// This error should only occur on metacluster due to the multi-stage process.
// Having temporary tenants may exceed capacity, so we disallow the rename.
ASSERT(operationType == OperationType::METACLUSTER);
return Void();
} else {
try {
// In the case of commit_unknown_result, assume we continue retrying
// until it's successful. Next loop around may throw error because it's
// already been moved, so account for that and update internal map as needed.
if (e.code() == error_code_commit_unknown_result) {
ASSERT(operationType != OperationType::MANAGEMENT_DATABASE &&
operationType != OperationType::METACLUSTER);
unknownResult = true;
}
wait(tr->onError(e));
} catch (Error& e) {
TraceEvent(SevError, "RenameTenantFailure")
.error(e)
.detail("TenantRenames", describe(tenantRenames));
return Void();
}
}
}
}
}
// Changes the configuration of a tenant
ACTOR static Future<Void> configureTenantImpl(Reference<ReadYourWritesTransaction> tr,
TenantName tenant,
std::map<Standalone<StringRef>, Optional<Value>> configParameters,
OperationType operationType,
bool specialKeysUseInvalidTuple,
TenantManagementWorkload* self) {
if (operationType == OperationType::SPECIAL_KEYS) {
tr->setOption(FDBTransactionOptions::SPECIAL_KEY_SPACE_ENABLE_WRITES);
for (auto const& [config, value] : configParameters) {
Tuple t;
if (specialKeysUseInvalidTuple) {
// Wrong number of items
if (deterministicRandom()->coinflip()) {
int numItems = deterministicRandom()->randomInt(0, 3);
if (numItems > 0) {
t.append(tenant);
}
if (numItems > 1) {
t.append(config).append(""_sr);
}
}
// Wrong data types
else {
if (deterministicRandom()->coinflip()) {
t.append(0).append(config);
} else {
t.append(tenant).append(0);
}
}
} else {
t.append(tenant).append(config);
}
if (value.present()) {
tr->set(self->specialKeysTenantConfigPrefix.withSuffix(t.pack()), value.get());
} else {
tr->clear(self->specialKeysTenantConfigPrefix.withSuffix(t.pack()));
}
}
wait(tr->commit());
ASSERT(!specialKeysUseInvalidTuple);
} else if (operationType == OperationType::METACLUSTER) {
wait(MetaclusterAPI::configureTenant(self->mvDb, tenant, configParameters));
} else {
// We don't have a transaction or database variant of this function
ASSERT(false);
}
return Void();
}
ACTOR static Future<Void> configureTenant(TenantManagementWorkload* self) {
state OperationType operationType =
deterministicRandom()->coinflip() ? OperationType::SPECIAL_KEYS : OperationType::METACLUSTER;
state TenantName tenant = self->chooseTenantName(true);
auto itr = self->createdTenants.find(tenant);
state bool exists = itr != self->createdTenants.end();
state Reference<ReadYourWritesTransaction> tr = makeReference<ReadYourWritesTransaction>(self->dataDb);
state std::map<Standalone<StringRef>, Optional<Value>> configuration;
state Optional<TenantGroupName> newTenantGroup;
// If true, the options generated may include an unknown option
state bool hasInvalidOption = deterministicRandom()->random01() < 0.1;
// True if any tenant group name starts with \xff
state bool hasSystemTenantGroup = false;
state bool specialKeysUseInvalidTuple =
operationType == OperationType::SPECIAL_KEYS && deterministicRandom()->random01() < 0.1;
// Generate a tenant group. Sometimes do this at the same time that we include an invalid option to ensure
// that the configure function still fails
if (!hasInvalidOption || deterministicRandom()->coinflip()) {
newTenantGroup = self->chooseTenantGroup(true);
hasSystemTenantGroup = hasSystemTenantGroup || newTenantGroup.orDefault(""_sr).startsWith("\xff"_sr);
configuration["tenant_group"_sr] = newTenantGroup;
}
if (hasInvalidOption) {
configuration["invalid_option"_sr] = ""_sr;
}
loop {
try {
wait(configureTenantImpl(tr, tenant, configuration, operationType, specialKeysUseInvalidTuple, self));
ASSERT(exists);
ASSERT(!hasInvalidOption);
ASSERT(!hasSystemTenantGroup);
ASSERT(!specialKeysUseInvalidTuple);
auto itr = self->createdTenants.find(tenant);
if (itr->second.tenantGroup.present()) {
auto tenantGroupItr = self->createdTenantGroups.find(itr->second.tenantGroup.get());
ASSERT(tenantGroupItr != self->createdTenantGroups.end());
if (--tenantGroupItr->second.tenantCount == 0) {
self->createdTenantGroups.erase(tenantGroupItr);
}
}
if (newTenantGroup.present()) {
self->createdTenantGroups[newTenantGroup.get()].tenantCount++;
}
itr->second.tenantGroup = newTenantGroup;
return Void();
} catch (Error& e) {
state Error error = e;
if (e.code() == error_code_tenant_not_found) {
ASSERT(!exists);
return Void();
} else if (e.code() == error_code_special_keys_api_failure) {
ASSERT(hasInvalidOption || specialKeysUseInvalidTuple);
return Void();
} else if (e.code() == error_code_invalid_tenant_configuration) {
ASSERT(hasInvalidOption);
return Void();
} else if (e.code() == error_code_invalid_metacluster_operation) {
ASSERT(operationType == OperationType::METACLUSTER != self->useMetacluster);
return Void();
} else if (e.code() == error_code_invalid_tenant_group_name) {
ASSERT(hasSystemTenantGroup);
return Void();
}
try {
wait(tr->onError(e));
} catch (Error&) {
TraceEvent(SevError, "ConfigureTenantFailure").error(error).detail("TenantName", tenant);
return Void();
}
}
}
}
// Gets the metadata for a tenant group using the specified operation type
ACTOR static Future<Optional<TenantGroupEntry>> getTenantGroupImpl(Reference<ReadYourWritesTransaction> tr,
TenantGroupName tenant,
OperationType operationType,
TenantManagementWorkload* self) {
state Optional<TenantGroupEntry> entry;
if (operationType == OperationType::MANAGEMENT_DATABASE) {
wait(store(entry, TenantAPI::tryGetTenantGroup(self->dataDb.getReference(), tenant)));
} else if (operationType == OperationType::MANAGEMENT_TRANSACTION ||
operationType == OperationType::SPECIAL_KEYS) {
// There is no special-keys interface for reading tenant groups currently, so read them
// using the TenantAPI.
tr->setOption(FDBTransactionOptions::READ_SYSTEM_KEYS);
wait(store(entry, TenantAPI::tryGetTenantGroupTransaction(tr, tenant)));
} else {
wait(store(entry, MetaclusterAPI::tryGetTenantGroup(self->mvDb, tenant)));
}
return entry;
}
ACTOR static Future<Void> getTenantGroup(TenantManagementWorkload* self) {
state TenantGroupName tenantGroup = self->chooseTenantGroup(true, false).get();
state OperationType operationType = self->randomOperationType();
state Reference<ReadYourWritesTransaction> tr = makeReference<ReadYourWritesTransaction>(self->dataDb);
// True if the tenant group should should exist and return a result
auto itr = self->createdTenantGroups.find(tenantGroup);
state bool alreadyExists = itr != self->createdTenantGroups.end() &&
!(operationType == OperationType::METACLUSTER && !self->useMetacluster);
loop {
try {
// Get the tenant group metadata and check that it matches our local state
state Optional<TenantGroupEntry> entry = wait(getTenantGroupImpl(tr, tenantGroup, operationType, self));
ASSERT(alreadyExists == entry.present());
if (entry.present()) {
ASSERT(entry.get().assignedCluster.present() == (operationType == OperationType::METACLUSTER));
}
return Void();
} catch (Error& e) {
state bool retry = false;
state Error error = e;
// Transaction-based operations should retry
if (operationType == OperationType::MANAGEMENT_TRANSACTION ||
operationType == OperationType::SPECIAL_KEYS) {
try {
wait(tr->onError(e));
retry = true;
} catch (Error& e) {
error = e;
retry = false;
}
}
if (!retry) {
TraceEvent(SevError, "GetTenantGroupFailure").error(error).detail("TenantGroupName", tenantGroup);
return Void();
}
}
}
}
// Gets a list of tenant groups using the specified operation type
ACTOR static Future<std::vector<std::pair<TenantGroupName, TenantGroupEntry>>> listTenantGroupsImpl(
Reference<ReadYourWritesTransaction> tr,
TenantGroupName beginTenantGroup,
TenantGroupName endTenantGroup,
int limit,
OperationType operationType,
TenantManagementWorkload* self) {
state std::vector<std::pair<TenantGroupName, TenantGroupEntry>> tenantGroups;
if (operationType == OperationType::MANAGEMENT_DATABASE) {
wait(store(
tenantGroups,
TenantAPI::listTenantGroups(self->dataDb.getReference(), beginTenantGroup, endTenantGroup, limit)));
} else if (operationType == OperationType::MANAGEMENT_TRANSACTION ||
operationType == OperationType::SPECIAL_KEYS) {
tr->setOption(FDBTransactionOptions::READ_SYSTEM_KEYS);
wait(store(tenantGroups,
TenantAPI::listTenantGroupsTransaction(tr, beginTenantGroup, endTenantGroup, limit)));
} else {
wait(store(tenantGroups,
MetaclusterAPI::listTenantGroups(self->mvDb, beginTenantGroup, endTenantGroup, limit)));
}
return tenantGroups;
}
ACTOR static Future<Void> listTenantGroups(TenantManagementWorkload* self) {
state TenantGroupName beginTenantGroup = self->chooseTenantGroup(false, false).get();
state TenantGroupName endTenantGroup = self->chooseTenantGroup(false, false).get();
state int limit = std::min(CLIENT_KNOBS->MAX_TENANTS_PER_CLUSTER + 1,
deterministicRandom()->randomInt(1, self->maxTenants * 2));
state OperationType operationType = self->randomOperationType();
state Reference<ReadYourWritesTransaction> tr = makeReference<ReadYourWritesTransaction>(self->dataDb);
if (beginTenantGroup > endTenantGroup) {
std::swap(beginTenantGroup, endTenantGroup);
}
loop {
try {
// Attempt to read the chosen list of tenant groups
state std::vector<std::pair<TenantGroupName, TenantGroupEntry>> tenantGroups =
wait(listTenantGroupsImpl(tr, beginTenantGroup, endTenantGroup, limit, operationType, self));
// Attempting to read the list of tenant groups using the metacluster API in a non-metacluster should
// return nothing in this test
if (operationType == OperationType::METACLUSTER && !self->useMetacluster) {
ASSERT(tenantGroups.size() == 0);
return Void();
}
ASSERT(tenantGroups.size() <= limit);
// Compare the resulting tenant list to the list we expected to get
auto localItr = self->createdTenantGroups.lower_bound(beginTenantGroup);
auto tenantMapItr = tenantGroups.begin();
for (; tenantMapItr != tenantGroups.end(); ++tenantMapItr, ++localItr) {
ASSERT(localItr != self->createdTenantGroups.end());
ASSERT(localItr->first == tenantMapItr->first);
}
// Make sure the list terminated at the right spot
ASSERT(tenantGroups.size() == limit || localItr == self->createdTenantGroups.end() ||
localItr->first >= endTenantGroup);
return Void();
} catch (Error& e) {
state bool retry = false;
state Error error = e;
// Transaction-based operations need to be retried
if (operationType == OperationType::MANAGEMENT_TRANSACTION ||
operationType == OperationType::SPECIAL_KEYS) {
try {
retry = true;
wait(tr->onError(e));
} catch (Error& e) {
error = e;
retry = false;
}
}
if (!retry) {
TraceEvent(SevError, "ListTenantGroupFailure")
.error(error)
.detail("BeginTenant", beginTenantGroup)
.detail("EndTenant", endTenantGroup);
return Void();
}
}
}
}
Future<Void> start(Database const& cx) override {
if (clientId == 0 || !singleClient) {
return _start(cx, this);
} else {
return Void();
}
}
ACTOR Future<Void> _start(Database cx, TenantManagementWorkload* self) {
state double start = now();
// Run a random sequence of tenant management operations for the duration of the test
while (now() < start + self->testDuration) {
state int operation = deterministicRandom()->randomInt(0, 8);
if (operation == 0) {
wait(createTenant(self));
} else if (operation == 1) {
wait(deleteTenant(self));
} else if (operation == 2) {
wait(getTenant(self));
} else if (operation == 3) {
wait(listTenants(self));
} else if (operation == 4) {
wait(renameTenant(self));
} else if (operation == 5) {
wait(configureTenant(self));
} else if (operation == 6) {
wait(getTenantGroup(self));
} else if (operation == 7) {
wait(listTenantGroups(self));
}
}
return Void();
}
// Verify that the set of tenants in the database matches our local state
ACTOR static Future<Void> compareTenants(TenantManagementWorkload* self) {
state std::map<TenantName, TenantData>::iterator localItr = self->createdTenants.begin();
state std::vector<Future<Void>> checkTenants;
state TenantName beginTenant = ""_sr.withPrefix(self->localTenantNamePrefix);
state TenantName endTenant = "\xff\xff"_sr.withPrefix(self->localTenantNamePrefix);
loop {
// Read the tenant list from the data cluster.
state std::vector<std::pair<TenantName, TenantMapEntry>> dataClusterTenants =
wait(TenantAPI::listTenants(self->dataDb.getReference(), beginTenant, endTenant, 1000));
auto dataItr = dataClusterTenants.begin();
TenantNameRef lastTenant;
while (dataItr != dataClusterTenants.end()) {
ASSERT(localItr != self->createdTenants.end());
ASSERT(dataItr->first == localItr->first);
ASSERT(dataItr->second.tenantGroup == localItr->second.tenantGroup);
ASSERT(dataItr->second.encrypted == localItr->second.encrypted);
checkTenants.push_back(checkTenantContents(self, dataItr->first, localItr->second));
lastTenant = dataItr->first;
++localItr;
++dataItr;
}
if (dataClusterTenants.size() < 1000) {
break;
} else {
beginTenant = keyAfter(lastTenant);
}
}
ASSERT(localItr == self->createdTenants.end());
wait(waitForAll(checkTenants));
return Void();
}
// Check that the given tenant group has the expected number of tenants
ACTOR template <class DB>
static Future<Void> checkTenantGroupTenantCount(Reference<DB> db, TenantGroupName tenantGroup, int expectedCount) {
TenantGroupName const& tenantGroupRef = tenantGroup;
int const& expectedCountRef = expectedCount;
KeyBackedSet<Tuple>::RangeResultType tenants =
wait(runTransaction(db, [tenantGroupRef, expectedCountRef](Reference<typename DB::TransactionT> tr) {
tr->setOption(FDBTransactionOptions::READ_SYSTEM_KEYS);
return TenantMetadata::tenantGroupTenantIndex().getRange(tr,
Tuple::makeTuple(tenantGroupRef),
Tuple::makeTuple(keyAfter(tenantGroupRef)),
expectedCountRef + 1);
}));
ASSERT(tenants.results.size() == expectedCount && !tenants.more);
return Void();
}
// Verify that the set of tenants in the database matches our local state
ACTOR static Future<Void> compareTenantGroups(TenantManagementWorkload* self) {
state std::map<TenantName, TenantGroupData>::iterator localItr = self->createdTenantGroups.begin();
state TenantName beginTenantGroup = ""_sr.withPrefix(self->localTenantGroupNamePrefix);
state TenantName endTenantGroup = "\xff\xff"_sr.withPrefix(self->localTenantGroupNamePrefix);
state std::vector<Future<Void>> checkTenantGroups;
loop {
// Read the tenant group list from the data cluster.
state KeyBackedRangeResult<std::pair<TenantGroupName, TenantGroupEntry>> dataClusterTenantGroups;
TenantName const& beginTenantGroupRef = beginTenantGroup;
TenantName const& endTenantGroupRef = endTenantGroup;
wait(
store(dataClusterTenantGroups,
runTransaction(self->dataDb.getReference(),
[beginTenantGroupRef, endTenantGroupRef](Reference<ReadYourWritesTransaction> tr) {
tr->setOption(FDBTransactionOptions::READ_SYSTEM_KEYS);
return TenantMetadata::tenantGroupMap().getRange(
tr, beginTenantGroupRef, endTenantGroupRef, 1000);
})));
auto dataItr = dataClusterTenantGroups.results.begin();
TenantGroupNameRef lastTenantGroup;
while (dataItr != dataClusterTenantGroups.results.end()) {
ASSERT(localItr != self->createdTenantGroups.end());
ASSERT(dataItr->first == localItr->first);
ASSERT(!dataItr->second.assignedCluster.present());
lastTenantGroup = dataItr->first;
checkTenantGroups.push_back(checkTenantGroupTenantCount(
self->dataDb.getReference(), dataItr->first, localItr->second.tenantCount));
++localItr;
++dataItr;
}
if (!dataClusterTenantGroups.more) {
break;
} else {
beginTenantGroup = keyAfter(lastTenantGroup);
}
}
ASSERT(localItr == self->createdTenantGroups.end());
return Void();
}
// Check that the tenant tombstones are properly cleaned up
ACTOR static Future<Void> checkTombstoneCleanup(TenantManagementWorkload* self) {
state Reference<ReadYourWritesTransaction> tr = self->dataDb->createTransaction();
loop {
try {
tr->setOption(FDBTransactionOptions::READ_SYSTEM_KEYS);
Optional<TenantTombstoneCleanupData> tombstoneCleanupData =
wait(TenantMetadata::tombstoneCleanupData().get(tr));
if (self->oldestDeletionVersion != 0) {
ASSERT(tombstoneCleanupData.present());
if (self->newestDeletionVersion - self->oldestDeletionVersion >
CLIENT_KNOBS->TENANT_TOMBSTONE_CLEANUP_INTERVAL * CLIENT_KNOBS->VERSIONS_PER_SECOND) {
ASSERT(tombstoneCleanupData.get().tombstonesErasedThrough >= 0);
}
}
break;
} catch (Error& e) {
wait(tr->onError(e));
}
}
return Void();
}
Future<bool> check(Database const& cx) override {
if (clientId == 0 || !singleClient) {
return _check(cx, this);
} else {
return true;
}
}
ACTOR static Future<bool> _check(Database cx, TenantManagementWorkload* self) {
state Transaction tr(self->dataDb);
// Check that the key we set outside of the tenant is present and has the correct value
// This is the same key we set inside some of our tenants, so this checks that no tenant
// writes accidentally happened in the raw key-space
loop {
try {
tr.setOption(FDBTransactionOptions::RAW_ACCESS);
Optional<Value> val = wait(tr.get(self->keyName));
ASSERT(val.present() && val.get() == self->noTenantValue);
break;
} catch (Error& e) {
wait(tr.onError(e));
}
}
wait(compareTenants(self) && compareTenantGroups(self));
if (self->useMetacluster) {
// The metacluster consistency check runs the tenant consistency check for each cluster
state MetaclusterConsistencyCheck<IDatabase> metaclusterConsistencyCheck(
self->mvDb, AllowPartialMetaclusterOperations::False);
wait(metaclusterConsistencyCheck.run());
wait(checkTombstoneCleanup(self));
} else {
state TenantConsistencyCheck<DatabaseContext> tenantConsistencyCheck(self->dataDb.getReference());
wait(tenantConsistencyCheck.run());
}
return true;
}
void getMetrics(std::vector<PerfMetric>& m) override {}
};
WorkloadFactory<TenantManagementWorkload> TenantManagementWorkload("TenantManagement");