Hawkin
/
arangodb
mirror of https://gitee.com/bigwinds/arangodb
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Merge branch 'generic-col-types' of github.com:arangodb/arangodb into obi-velocystream 

* 'generic-col-types' of github.com:arangodb/arangodb:
  Improve version handling in ClusterInfo.
  Add further logging where a bug is likely.
  make error message unambiguous

Conflicts:
	lib/Rest/HttpResponse.cpp
This commit is contained in:
Jan Christoph Uhde 2016-09-05 16:42:41 +02:00
parent 4c473753b7 49d7cf345f
commit d0fdfa727a
5 changed files with 73 additions and 55 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

50
arangod/Cluster/ClusterInfo.cpp
View File

@ -379,9 +379,13 @@ void ClusterInfo::loadPlan() {
DatabaseFeature* databaseFeature =
application_features::ApplicationServer::getFeature<DatabaseFeature>(
"Database");
uint64_t storedVersion = _planProt.version;
MUTEX_LOCKER(mutexLocker, _planProt.mutex);
if (_planProt.version > storedVersion) {
++_planProt.wantedVersion; // Indicate that after *NOW* somebody has to
// reread from the agency!
MUTEX_LOCKER(mutexLocker, _planProt.mutex); // only one may work at a time
uint64_t storedVersion = _planProt.wantedVersion; // this is the version
// we will set in the end
if (_planProt.doneVersion == storedVersion) {
// Somebody else did, what we intended to do, so just return
return;
}
@ -502,7 +506,7 @@ void ClusterInfo::loadPlan() {
_shards.swap(newShards);
_shardKeys.swap(newShardKeys);
}
_planProt.version++; // such that others notice our change
_planProt.doneVersion = storedVersion;
_planProt.isValid = true; // will never be reset to false
} else {
LOG(ERR) << "\"Plan\" is not an object in agency";
@ -525,9 +529,12 @@ void ClusterInfo::loadPlan() {
static std::string const prefixCurrent = "Current";
void ClusterInfo::loadCurrent() {
uint64_t storedVersion = _currentProt.version;
MUTEX_LOCKER(mutexLocker, _currentProt.mutex);
if (_currentProt.version > storedVersion) {
++_currentProt.wantedVersion; // Indicate that after *NOW* somebody has to
// reread from the agency!
MUTEX_LOCKER(mutexLocker, _currentProt.mutex); // only one may work at a time
uint64_t storedVersion = _currentProt.wantedVersion; // this is the version
// we will set at the end
if (_currentProt.doneVersion == storedVersion) {
// Somebody else did, what we intended to do, so just return
return;
}
@ -617,7 +624,7 @@ void ClusterInfo::loadCurrent() {
_currentCollections.swap(newCollections);
_shardIds.swap(newShardIds);
}
_currentProt.version++; // such that others notice our change
_currentProt.doneVersion = storedVersion;
_currentProt.isValid = true; // will never be reset to false
} else {
LOG(ERR) << "Current is not an object!";
@ -1851,9 +1858,12 @@ static std::string const prefixServers = "Current/ServersRegistered";
void ClusterInfo::loadServers() {
uint64_t storedVersion = _serversProt.version;
++_serversProt.wantedVersion; // Indicate that after *NOW* somebody has to
// reread from the agency!
MUTEX_LOCKER(mutexLocker, _serversProt.mutex);
if (_serversProt.version > storedVersion) {
uint64_t storedVersion = _serversProt.wantedVersion; // this is the version
// we will set in the end
if (_serversProt.doneVersion == storedVersion) {
// Somebody else did, what we intended to do, so just return
return;
}
@ -1883,7 +1893,7 @@ void ClusterInfo::loadServers() {
{
WRITE_LOCKER(writeLocker, _serversProt.lock);
_servers.swap(newServers);
_serversProt.version++; // such that others notice our change
_serversProt.doneVersion = storedVersion;
_serversProt.isValid = true; // will never be reset to false
}
return;
@ -1976,9 +1986,12 @@ std::string ClusterInfo::getServerName(std::string const& endpoint) {
static std::string const prefixCurrentCoordinators = "Current/Coordinators";
void ClusterInfo::loadCurrentCoordinators() {
uint64_t storedVersion = _coordinatorsProt.version;
++_coordinatorsProt.wantedVersion; // Indicate that after *NOW* somebody
// has to reread from the agency!
MUTEX_LOCKER(mutexLocker, _coordinatorsProt.mutex);
if (_coordinatorsProt.version > storedVersion) {
uint64_t storedVersion = _coordinatorsProt.wantedVersion; // this is the
// version we will set in the end
if (_coordinatorsProt.doneVersion == storedVersion) {
// Somebody else did, what we intended to do, so just return
return;
}
@ -2004,7 +2017,7 @@ void ClusterInfo::loadCurrentCoordinators() {
{
WRITE_LOCKER(writeLocker, _coordinatorsProt.lock);
_coordinators.swap(newCoordinators);
_coordinatorsProt.version++; // such that others notice our change
_coordinatorsProt.doneVersion = storedVersion;
_coordinatorsProt.isValid = true; // will never be reset to false
}
return;
@ -2028,9 +2041,12 @@ static std::string const prefixTargetCleaned = "Target/CleanedOutServers";
static std::string const prefixTargetFailed = "Target/FailedServers";
void ClusterInfo::loadCurrentDBServers() {
uint64_t storedVersion = _DBServersProt.version;
++_DBServersProt.wantedVersion; // Indicate that after *NOW* somebody has to
// reread from the agency!
MUTEX_LOCKER(mutexLocker, _DBServersProt.mutex);
if (_DBServersProt.version > storedVersion) {
uint64_t storedVersion = _DBServersProt.wantedVersion; // this is the version
// we will set in the end
if (_DBServersProt.doneVersion == storedVersion) {
// Somebody else did, what we intended to do, so just return
return;
}
@ -2089,7 +2105,7 @@ void ClusterInfo::loadCurrentDBServers() {
{
WRITE_LOCKER(writeLocker, _DBServersProt.lock);
_DBServers.swap(newDBServers);
_DBServersProt.version++; // such that others notice our change
_DBServersProt.doneVersion = storedVersion;
_DBServersProt.isValid = true; // will never be reset to false
}
return;

26
arangod/Cluster/ClusterInfo.h
View File

@ -529,28 +529,28 @@ class ClusterInfo {
// Cached data from the agency, we reload whenever necessary:
// We group the data, each group has an atomic "valid-flag"
// which is used for lazy loading in the beginning. It starts
// as false, is set to true at each reload and is never reset
// to false in the lifetime of the server. The variable is
// atomic to be able to check it without acquiring
// the read lock (see below). Flush is just an explicit reload
// for all data and is only used in tests.
// We group the data, each group has an atomic "valid-flag" which is
// used for lazy loading in the beginning. It starts as false, is set
// to true at each reload and is only reset to false if the cache
// needs to be invalidated. The variable is atomic to be able to check
// it without acquiring the read lock (see below). Flush is just an
// explicit reload for all data and is only used in tests.
// Furthermore, each group has a mutex that protects against
// simultaneously contacting the agency for an update.
// In addition, each group has an atomic version number, this is used
// to prevent a stampede if multiple threads notice concurrently
// that an update from the agency is necessary. Finally, there is
// a read/write lock which protects the actual data structure.
// In addition, each group has two atomic version numbers, these are
// used to prevent a stampede if multiple threads notice concurrently
// that an update from the agency is necessary. Finally, there is a
// read/write lock which protects the actual data structure.
// We encapsulate this protection in the struct ProtectionData:
struct ProtectionData {
std::atomic<bool> isValid;
Mutex mutex;
std::atomic<uint64_t> version;
std::atomic<uint64_t> wantedVersion;
std::atomic<uint64_t> doneVersion;
arangodb::basics::ReadWriteLock lock;
ProtectionData() : isValid(false), version(0) {}
ProtectionData() : isValid(false), wantedVersion(0), doneVersion(0) {}
};
// The servers, first all, we only need Current here:

6
arangod/Replication/InitialSyncer.cpp
View File

@ -699,8 +699,7 @@ int InitialSyncer::handleCollectionDump(
}
if (response->getHttpReturnCode() == 404) {
// unknown job, we can abort
errorMsg = "no response received from master at " +
_masterInfo._endpoint;
errorMsg = "job not found on master at " + _masterInfo._endpoint;
return TRI_ERROR_REPLICATION_NO_RESPONSE;
}
}
@ -889,8 +888,7 @@ int InitialSyncer::handleCollectionSync(
}
if (response->getHttpReturnCode() == 404) {
// unknown job, we can abort
errorMsg = "no response received from master at " +
_masterInfo._endpoint;
errorMsg = "job not found on master at " + _masterInfo._endpoint;
return TRI_ERROR_REPLICATION_NO_RESPONSE;
}
}

7
lib/Rest/GeneralResponse.h
View File

@ -141,7 +141,12 @@ class GeneralResponse {
addPayload(std::forward<Payload>(payload), &options, resolveExternals);
}
void addPayloadPreconditions() { TRI_ASSERT(_vpackPayloads.size() == 0); }
void addPayloadPreconditions() {
if (_vpackPayloads.size() != 0) {
LOG(ERR) << "Payload set twice";
TRI_ASSERT(_vpackPayloads.size() == 0);
}
}
virtual void addPayloadPreHook(bool inputIsBuffer, bool& resolveExternals,
bool& skipBody) {}
void addPayload(VPackSlice const&,

1
lib/Rest/HttpResponse.cpp
View File

@ -306,7 +306,6 @@ void HttpResponse::addPayloadPostHook(
bool resolveExternals = true, bool bodySkipped = false) {
VPackSlice const* slicePtr;
VPackSlice tmpSlice;
if (!bodySkipped) {
// we have Probably resolved externals
TRI_ASSERT(!_vpackPayloads.empty());