//! メンバ追加のテスト（複数）
TEST_F( ClusterInfoTest, test_addMembers )
{
ClusterInfo ci;
const vector< DataIndex >* pMembers;
vector< DataIndex > addedData;
addedData.push_back( 5 );
addedData.push_back( 7 );
addedData.push_back( 9 );
ci.addMembers( addedData );
pMembers = ci.getMembers();
ASSERT_EQ( 3, pMembers->size() );
EXPECT_EQ( 5, pMembers->at( 0 ) );
EXPECT_EQ( 7, pMembers->at( 1 ) );
EXPECT_EQ( 9, pMembers->at( 2 ) );
addedData.clear();
addedData.push_back( 4 );
addedData.push_back( 6 );
addedData.push_back( 8 );
addedData.push_back( 10 );
ci.addMembers( addedData );
pMembers = ci.getMembers();
ASSERT_EQ( 7, pMembers->size() );
EXPECT_EQ( 5, pMembers->at( 0 ) ); // ※ソートしない
EXPECT_EQ( 7, pMembers->at( 1 ) );
EXPECT_EQ( 9, pMembers->at( 2 ) );
EXPECT_EQ( 4, pMembers->at( 3 ) );
EXPECT_EQ( 6, pMembers->at( 4 ) );
EXPECT_EQ( 8, pMembers->at( 5 ) );
EXPECT_EQ( 10, pMembers->at( 6 ) );
}
TRI_voc_cid_t CollectionNameResolver::getCollectionIdCluster(
std::string const& name) const {
if (!ServerState::isRunningInCluster(_serverRole)) {
return getCollectionIdLocal(name);
}
if (name[0] >= '0' && name[0] <= '9') {
// name is a numeric id
TRI_voc_cid_t cid = static_cast<TRI_voc_cid_t>(arangodb::basics::StringUtils::uint64(name));
// Now validate the cid
TRI_col_type_e type = getCollectionTypeCluster(getCollectionNameCluster(cid));
if (type == TRI_COL_TYPE_UNKNOWN) {
return 0;
}
return cid;
}
try {
// We have to look up the collection info:
ClusterInfo* ci = ClusterInfo::instance();
auto cinfo = ci->getCollection(_vocbase->name(), name);
TRI_ASSERT(cinfo != nullptr);
return cinfo->cid();
} catch (...) {
}
return 0;
}
/*
* Aim: to create binary tree of unique record id -> unique inventor id,
* also to allow fast search and insertion/deletion.
* the unique inventor id is also a const Record pointer,
* meaning that different unique record ids may be associated with a same
* const Record pointer that represents them.
*
* Algorithm: clean the uinv2count and uid2uinv tree first.
* For any cluster in the cCluser_Info object:
* For any const Record pointer p in the cluster member list:
* create a std::pair of (p, d), where d is the delegate
* (or representative) of the cluster
* insert the pair into uid2uinv map.
* End for
* End for
*
* uinv2count is updated in the same way.
*/
void
cBlocking_Operation_By_Coauthors::build_uid2uinv_tree(const ClusterInfo & cluster) {
uinv2count_tree.clear();
uid2uinv_tree.clear();
uint32_t count = 0;
//typedef list<Cluster> cRecGroup;
// Maybe should be RecordGroup
typedef list<Cluster> ClusterList;
std::cout << "Building trees: 1. Unique Record ID to Unique Inventer ID. ";
std::cout << "2 Unique Inventer ID to Number of holding patents ........";
std::cout << std::endl;
map<string, ClusterList>::const_iterator p = cluster.get_cluster_map().begin();
for (; p != cluster.get_cluster_map().end(); ++p) {
ClusterList::const_iterator q = p->second.begin();
for (; q != p->second.end(); ++q) {
const Record * value = q->get_cluster_head().m_delegate;
map<const Record *, uint32_t>::iterator pcount = uinv2count_tree.find(value);
if (pcount == uinv2count_tree.end())
pcount = uinv2count_tree.insert(std::pair<const Record *, uint32_t>(value, 0)).first;
for (RecordPList::const_iterator r = q->get_fellows().begin(); r != q->get_fellows().end(); ++r) {
const Record * key = *r;
uid2uinv_tree.insert(std::pair<const Record * , const Record *>(key, value ));
++(pcount->second);
++count;
}
}
}
std::cout << count << " nodes has been created inside the tree." << std::endl;
}
//! コンストラクタ（マージ）のテスト
TEST_F( ClusterInfoTest, test_mergeConstructor )
{
ClusterInfo *pC1 = new ClusterInfo();
pC1->addMember( 3 );
pC1->addMember( 5 );
pC1->addMember( 1 );
ClusterInfo *pC2 = new ClusterInfo();
pC2->addMember( 2 );
pC2->addMember( 4 );
pC2->addMember( 6 );
const vector< DataIndex >* pMembers = NULL;
ClusterInfo *pCm = new ClusterInfo( pC1, pC2, 100.0f, 123 );
pMembers = pCm->getMembers();
ASSERT_EQ( 6, pMembers->size() );
EXPECT_EQ( 1, pMembers->at( 0 ) ); // ※ソートされる
EXPECT_EQ( 2, pMembers->at( 1 ) );
EXPECT_EQ( 3, pMembers->at( 2 ) );
EXPECT_EQ( 4, pMembers->at( 3 ) );
EXPECT_EQ( 5, pMembers->at( 4 ) );
EXPECT_EQ( 6, pMembers->at( 5 ) );
ASSERT_EQ( 100.0f, pCm->getMergeCost() );
ASSERT_EQ( 123, pCm->getId() );
delete pCm;
delete pC1;
delete pC2;
return;
}
//! メンバ追加のテスト（複数）
TEST_F( ClusterInfoTest, test_setAndGetmergeCost )
{
ClusterInfo ci;
ci.setMergeCost( 0.5f );
EXPECT_EQ( 0.5f, ci.getMergeCost() );
ci.setMergeCost( 1234.567f );
EXPECT_EQ( 1234.567f, ci.getMergeCost() );
}
Float HierarchicalClustering::computeClusterVariance( const ClusterInfo &cluster, const MatrixFloat &data ){
VectorFloat mean(N,0);
VectorFloat std(N,0);
//Compute the mean
UINT numSamples = cluster.getNumSamplesInCluster();
for(UINT j=0; j<N; j++){
for(UINT i=0; i<numSamples; i++){
UINT index = cluster[i];
mean[j] += data[ index ][j];
}
mean[j] /= Float( numSamples );
}
//Compute the std dev
for(UINT j=0; j<N; j++){
for(UINT i=0; i<numSamples; i++){
std[j] += grt_sqr( data[ cluster[i] ][j] - mean[j] );
}
std[j] = grt_sqrt( std[j] / Float( numSamples-1 ) );
}
Float variance = 0;
for(UINT j=0; j<N; j++){
variance += std[j];
}
return variance/N;
}
//! メンバ追加のテスト
TEST_F( ClusterInfoTest, test_addMember )
{
ClusterInfo ci;
const vector< DataIndex >* pMembers;
ci.addMember( 1 );
pMembers = ci.getMembers();
ASSERT_EQ( 1, pMembers->size() );
EXPECT_EQ( 1, pMembers->at( 0 ) );
ci.addMember( 100 );
pMembers = ci.getMembers();
ASSERT_EQ( 2, pMembers->size() );
EXPECT_EQ( 1, pMembers->at( 0 ) );
EXPECT_EQ( 100, pMembers->at( 1 ) );
}
Float HierarchicalClustering::computeClusterDistance( const ClusterInfo &clusterA, const ClusterInfo &clusterB ){
Float minDist = grt_numeric_limits< Float >::max();
const UINT numSamplesA = clusterA.getNumSamplesInCluster();
const UINT numSamplesB = clusterB.getNumSamplesInCluster();
//Find the minimum distance between the two clusters
for(UINT i=0; i<numSamplesA; i++){
for(UINT j=0; j<numSamplesB; j++){
if( distanceMatrix[ clusterA[i] ][ clusterB[j] ] < minDist ){
minDist = distanceMatrix[ clusterA[i] ][ clusterB[j] ];
}
}
}
return minDist;
}
bool GPFSHandler::setBasicInfo()
{
// do an initial polling here to get some cluster info
if(pollinghandler == NULL)
return false;
pollinghandler->refreshClusterRecipe();
MErrno err = M_OK;
ClusterInfo* clusterInfo = new ClusterInfo(&err);
pollinghandler->updateClusterInfo(clusterInfo);
char cname[30] = {0};
strcpy(cname,clusterInfo->getName());
cluster = cname;
delete clusterInfo;
procname = DAEMON_NAME;
char hname[50] = {0};
if(gethostname(hname,50) < 0)
return false;
hostname = hname;
return true;
}
TRI_col_type_e CollectionNameResolver::getCollectionTypeCluster(
std::string const& name) const {
// This fires in Single server case as well
if (!ServerState::isCoordinator(_serverRole)) {
return getCollectionType(name);
}
if (name[0] >= '0' && name[0] <= '9') {
// name is a numeric id
return getCollectionTypeCluster(
getCollectionName(static_cast<TRI_voc_cid_t>(
arangodb::basics::StringUtils::uint64(name))));
}
try {
// We have to look up the collection info:
ClusterInfo* ci = ClusterInfo::instance();
auto cinfo = ci->getCollection(_vocbase->name(), name);
TRI_ASSERT(cinfo != nullptr);
return cinfo->type();
} catch(...) {
}
return TRI_COL_TYPE_UNKNOWN;
}
void ClusterFeature::prepare() {
ServerState::instance()->setDataPath(_dataPath);
ServerState::instance()->setLogPath(_logPath);
ServerState::instance()->setArangodPath(_arangodPath);
ServerState::instance()->setDBserverConfig(_dbserverConfig);
ServerState::instance()->setCoordinatorConfig(_coordinatorConfig);
V8DealerFeature* v8Dealer =
ApplicationServer::getFeature<V8DealerFeature>("V8Dealer");
v8Dealer->defineDouble("SYS_DEFAULT_REPLICATION_FACTOR_SYSTEM",
_systemReplicationFactor);
// create callback registery
_agencyCallbackRegistry.reset(
new AgencyCallbackRegistry(agencyCallbacksPath()));
// Initialize ClusterInfo library:
ClusterInfo::createInstance(_agencyCallbackRegistry.get());
// initialize ConnectionManager library
httpclient::ConnectionManager::initialize();
// create an instance (this will not yet create a thread)
ClusterComm::instance();
AgencyFeature* agency =
application_features::ApplicationServer::getFeature<AgencyFeature>(
"Agency");
if (agency->isEnabled() || _enableCluster) {
// initialize ClusterComm library, must call initialize only once
ClusterComm::initialize();
auto authenticationFeature =
application_features::ApplicationServer::getFeature<AuthenticationFeature>(
"Authentication");
if (authenticationFeature->isEnabled() && !authenticationFeature->hasUserdefinedJwt()) {
LOG(FATAL) << "Cluster authentication enabled but jwt not set via command line. Please"
<< " provide --server.jwt-secret which is used throughout the cluster.";
FATAL_ERROR_EXIT();
}
}
// return if cluster is disabled
if (!_enableCluster) {
return;
}
ServerState::instance()->setClusterEnabled();
// register the prefix with the communicator
AgencyComm::setPrefix(_agencyPrefix);
for (size_t i = 0; i < _agencyEndpoints.size(); ++i) {
std::string const unified = Endpoint::unifiedForm(_agencyEndpoints[i]);
if (unified.empty()) {
LOG(FATAL) << "invalid endpoint '" << _agencyEndpoints[i]
<< "' specified for --cluster.agency-endpoint";
FATAL_ERROR_EXIT();
}
AgencyComm::addEndpoint(unified);
}
// Now either _myId is set properly or _myId is empty and _myLocalInfo and
// _myAddress are set.
if (!_myAddress.empty()) {
ServerState::instance()->setAddress(_myAddress);
}
// disable error logging for a while
ClusterComm::instance()->enableConnectionErrorLogging(false);
// perform an initial connect to the agency
std::string const endpoints = AgencyComm::getEndpointsString();
if (!AgencyComm::initialize()) {
LOG(FATAL) << "Could not connect to agency endpoints (" << endpoints << ")";
FATAL_ERROR_EXIT();
}
ServerState::instance()->setLocalInfo(_myLocalInfo);
if (!_myId.empty()) {
ServerState::instance()->setId(_myId);
}
if (!_myRole.empty()) {
ServerState::RoleEnum role = ServerState::stringToRole(_myRole);
if (role == ServerState::ROLE_SINGLE ||
role == ServerState::ROLE_UNDEFINED) {
LOG(FATAL) << "Invalid role provided. Possible values: PRIMARY, "
"SECONDARY, COORDINATOR";
FATAL_ERROR_EXIT();
}
if (!ServerState::instance()->registerWithRole(role)) {
LOG(FATAL) << "Couldn't register at agency.";
FATAL_ERROR_EXIT();
}
}
ServerState::RoleEnum role = ServerState::instance()->getRole();
if (role == ServerState::ROLE_UNDEFINED) {
// no role found
LOG(FATAL) << "unable to determine unambiguous role for server '" << _myId
<< "'. No role configured in agency (" << endpoints << ")";
FATAL_ERROR_EXIT();
}
if (role == ServerState::ROLE_SINGLE) {
LOG(FATAL) << "determined single-server role for server '" << _myId
<< "'. Please check the configurarion in the agency ("
<< endpoints << ")";
FATAL_ERROR_EXIT();
}
if (_myId.empty()) {
_myId = ServerState::instance()->getId(); // has been set by getRole!
}
// check if my-address is set
if (_myAddress.empty()) {
// no address given, now ask the agency for our address
_myAddress = ServerState::instance()->getAddress();
}
// if nonempty, it has already been set above
// If we are a coordinator, we wait until at least one DBServer is there,
// otherwise we can do very little, in particular, we cannot create
// any collection:
if (role == ServerState::ROLE_COORDINATOR) {
ClusterInfo* ci = ClusterInfo::instance();
double start = TRI_microtime();
while (true) {
LOG(INFO) << "Waiting for DBservers to show up...";
ci->loadCurrentDBServers();
std::vector<ServerID> DBServers = ci->getCurrentDBServers();
if (DBServers.size() >= 1 &&
(DBServers.size() > 1 || TRI_microtime() - start > 15.0)) {
LOG(INFO) << "Found " << DBServers.size() << " DBservers.";
break;
}
sleep(1);
};
}
if (_myAddress.empty()) {
LOG(FATAL) << "unable to determine internal address for server '" << _myId
<< "'. Please specify --cluster.my-address or configure the "
"address for this server in the agency.";
FATAL_ERROR_EXIT();
}
// now we can validate --cluster.my-address
std::string const unified = Endpoint::unifiedForm(_myAddress);
if (unified.empty()) {
LOG(FATAL) << "invalid endpoint '" << _myAddress
<< "' specified for --cluster.my-address";
FATAL_ERROR_EXIT();
}
}
bool HierarchicalClustering::train_(MatrixFloat &data){
trained = false;
clusters.clear();
distanceMatrix.clear();
if( data.getNumRows() == 0 || data.getNumCols() == 0 ){
return false;
}
//Set the rows and columns
M = data.getNumRows();
N = data.getNumCols();
//Build the distance matrix
distanceMatrix.resize(M,M);
//Build the distance matrix
for(UINT i=0; i<M; i++){
for(UINT j=0; j<M; j++){
if( i== j ) distanceMatrix[i][j] = grt_numeric_limits< Float >::max();
else{
distanceMatrix[i][j] = squaredEuclideanDistance(data[i], data[j]);
}
}
}
//Build the initial clusters, at the start each sample gets its own cluster
UINT uniqueClusterID = 0;
Vector< ClusterInfo > clusterData(M);
for(UINT i=0; i<M; i++){
clusterData[i].uniqueClusterID = uniqueClusterID++;
clusterData[i].addSampleToCluster(i);
}
trainingLog << "Starting clustering..." << std::endl;
//Create the first cluster level, each sample is it's own cluster
UINT level = 0;
ClusterLevel newLevel;
newLevel.level = level;
for(UINT i=0; i<M; i++){
newLevel.clusters.push_back( clusterData[i] );
}
clusters.push_back( newLevel );
//Move to level 1 and start the search
level++;
bool keepClustering = true;
while( keepClustering ){
//Find the closest two clusters within the cluster data
Float minDist = grt_numeric_limits< Float >::max();
Vector< Vector< UINT > > clusterPairs;
UINT K = (UINT)clusterData.size();
for(UINT i=0; i<K; i++){
for(UINT j=0; j<K; j++){
if( i != j ){
Float dist = computeClusterDistance( clusterData[i], clusterData[j] );
if( dist < minDist ){
minDist = dist;
Vector< UINT > clusterPair(2);
clusterPair[0] = i;
clusterPair[1] = j;
clusterPairs.clear();
clusterPairs.push_back( clusterPair );
}
}
}
}
if( minDist == grt_numeric_limits< Float >::max() ){
keepClustering = false;
warningLog << "train_(MatrixFloat &data) - Failed to find any cluster at level: " << level << std::endl;
return false;
}else{
//Merge the two closest clusters together and create a new level
ClusterLevel newLevel;
newLevel.level = level;
//Create the new cluster
ClusterInfo newCluster;
newCluster.uniqueClusterID = uniqueClusterID++;
const UINT numClusterPairs = clusterPairs.getSize();
for(UINT k=0; k<numClusterPairs; k++){
//Add all the samples in the first cluster to the new cluster
UINT numSamplesInClusterA = clusterData[ clusterPairs[k][0] ].getNumSamplesInCluster();
for(UINT i=0; i<numSamplesInClusterA; i++){
UINT index = clusterData[ clusterPairs[k][0] ][ i ];
newCluster.addSampleToCluster( index );
}
//Add all the samples in the second cluster to the new cluster
UINT numSamplesInClusterB = clusterData[ clusterPairs[k][1] ].getNumSamplesInCluster();
for(UINT i=0; i<numSamplesInClusterB; i++){
UINT index = clusterData[ clusterPairs[k][1] ][ i ];
newCluster.addSampleToCluster( index );
}
//Compute the cluster variance
newCluster.clusterVariance = computeClusterVariance( newCluster, data );
//Remove the two cluster pairs (so they will not be used in the next search
UINT idA = clusterData[ clusterPairs[k][0] ].getUniqueClusterID();
UINT idB = clusterData[ clusterPairs[k][1] ].getUniqueClusterID();
UINT numRemoved = 0;
Vector< ClusterInfo >::iterator iter = clusterData.begin();
while( iter != clusterData.end() ){
if( iter->getUniqueClusterID() == idA || iter->getUniqueClusterID() == idB ){
iter = clusterData.erase( iter );
if( ++numRemoved >= 2 ) break;
}else iter++;
}
}
//Add the merged cluster to the clusterData
clusterData.push_back( newCluster );
//Add the new level and cluster data to the main cluster buffer
newLevel.clusters.push_back( newCluster );
clusters.push_back( newLevel );
//Update the level
level++;
}
//Check to see if we should stop clustering
if( level >= M ){
keepClustering = false;
}
if( clusterData.size() == 0 ){
keepClustering = false;
}
trainingLog << "Cluster level: " << level << " Number of clusters: " << clusters.back().getNumClusters() << std::endl;
}
//Flag that the model is trained
trained = true;
//Setup the cluster labels
clusterLabels.resize(numClusters);
for(UINT i=0; i<numClusters; i++){
clusterLabels[i] = i+1;
}
clusterLikelihoods.resize(numClusters,0);
clusterDistances.resize(numClusters,0);
return true;
}
//! 子クラスタの取得、設定のテスト
TEST_F( ClusterInfoTest, test_setAndGetChildren )
{
ClusterInfo c;
ClusterInfo c1;
c1.addMember( 1 );
c1.addMember( 100 );
ClusterInfo c2;
c2.addMember( 10 );
c2.addMember( 11 );
c2.addMember( 12 );
// 初期状態
ASSERT_EQ( NULL, c.getChild1() );
ASSERT_EQ( NULL, c.getChild2() );
ASSERT_EQ( 0, c.getMembers()->size() );
// 子クラスタを設定
c.setChildren( &c1, &c2 );
ASSERT_EQ( &c1, c.getChild1() );
ASSERT_EQ( &c2, c.getChild2() );
ASSERT_EQ( 5, c.getMembers()->size() );
ASSERT_EQ( 1, c.getMembers()->at( 0 ) );
ASSERT_EQ( 10, c.getMembers()->at( 1 ) );
ASSERT_EQ( 11, c.getMembers()->at( 2 ) );
ASSERT_EQ( 12, c.getMembers()->at( 3 ) );
ASSERT_EQ( 100, c.getMembers()->at( 4 ) );
// 両方NULLを設定（メンバは変わらない）
c.setChildren( NULL, NULL );
ASSERT_EQ( NULL, c.getChild1() );
ASSERT_EQ( NULL, c.getChild2() );
ASSERT_EQ( 5, c.getMembers()->size() );
ASSERT_EQ( 1, c.getMembers()->at( 0 ) );
ASSERT_EQ( 10, c.getMembers()->at( 1 ) );
ASSERT_EQ( 11, c.getMembers()->at( 2 ) );
ASSERT_EQ( 12, c.getMembers()->at( 3 ) );
ASSERT_EQ( 100, c.getMembers()->at( 4 ) );
// 片方NULLのみを設定（メンバは変わる）
c.setChildren( NULL, &c1 );
ASSERT_EQ( NULL, c.getChild1() );
ASSERT_EQ( &c1, c.getChild2() );
ASSERT_EQ( 2, c.getMembers()->size() );
ASSERT_EQ( 1, c.getMembers()->at( 0 ) );
ASSERT_EQ( 100, c.getMembers()->at( 1 ) );
return;
}
std::string DistributeBlock::createKey(VPackSlice) const {
ClusterInfo* ci = ClusterInfo::instance();
uint64_t uid = ci->uniqid();
return std::to_string(uid);
}
void GPFSConfigHandler::task()
{
int nFSs = 0;
int nPools = 0;
int nDisks = 0;
int nFsets = 0;
int nNodes = 0;
TEAL_ERR_T ret;
string msg;
char tmp[10];
string fsName;
string stgName;
string diskName;
string fsetName;
string nodeName;
string clusterName;
FilesystemInfo* fsInfo = NULL;
StoragePoolInfo* stgInfo = NULL;
DiskInfo* diskInfo = NULL;
FileSet* fsetInfo = NULL;
FileSet* fileSetList = NULL;
MErrno err = M_OK;
log_info("########################Start refreshing all entities#########################################");
err = GPFSHandler::getPollHandler()->getDaemonState();
if(err != M_OK)
{
msg = "daemon is down on local node ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
err = GPFSHandler::getPollHandler()->refreshClusterRecipe();
if(err != M_OK)
{
msg = "refresh cluster failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
ClusterInfo* clusterInfo = new ClusterInfo(&err);
//update cluster info
err = GPFSHandler::getPollHandler()->updateClusterInfo(clusterInfo);
if(err != M_OK)
{
msg = "update cluster info failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
//update all nodes info
err = GPFSHandler::getPollHandler()->updateNodeInfo(clusterInfo);
if(err != M_OK)
{
msg = "update node failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
err = GPFSHandler::getPollHandler()->getClusterInfo(clusterInfo); //this maybe not needed
if(err != M_OK)
{
msg = "get cluster info failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
err = GPFSHandler::getPollHandler()->updateDiskSDRInfo();
if(err != M_OK)
{ /*TODO: This API invokes "mmsdrquery 30 3001:3004:3005:3006:3007:3008:3002:3003" under the cover. Need to check if it is a real error or an expected configuration to determin whether to ignore it or not.*/
msg = "update disk SDR info failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
msg += ", ignore it...";
log_warn(msg);
// return; // simply ignore it since there a configuration of two clusters and NSD may not be seen from the FS cluster.
}
err = GPFSHandler::getPollHandler()->updateFilesystemInfo(clusterInfo, 1);// to get perfermance statics even if not used.
if(err != M_OK)
{
msg = "update file system failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
err = GPFSHandler::getPollHandler()->updateMountedNodeInfo(clusterInfo); // to get mounted node info
if(err != M_OK)
{ /*TODO: This API invokes "mmlsmount all_local -Y" under the cover. Need to check if it is a real error or an expected configuration to determin whether to ignore it or not.*/
msg = "update mounted node info failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
msg += ", ignore it...";
log_warn(msg);
// return; // simply ignore it since there maybe no local file system configured
}
err = GPFSHandler::getPollHandler()->updateVfsStatsInfo(clusterInfo); // to get vfs info
if(err != M_OK)
{
msg = "update vfs info failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
err = GPFSHandler::getPollHandler()->updateThreadUtilInfo(clusterInfo); // to get thread util info
if(err != M_OK)
{
msg = "update thread util info failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
err = GPFSHandler::getPollHandler()->updateIocStatsInfo(clusterInfo); // to get ioc statics info
if(err != M_OK)
{
msg = "update ioc statics info failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
err = GPFSHandler::getPollHandler()->updateCacheStatsInfo(clusterInfo); // to get cache statics info
if(err != M_OK)
{
msg = "update cache statics info failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
err = GPFSHandler::getPollHandler()->updatePCacheStatsInfo(clusterInfo); // to get pcache statics info
if(err != M_OK)
{
msg = "update pcache statics info failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
err = GPFSHandler::getPollHandler()->updateFilesystemManagerInfo(clusterInfo);// update fs manager
if(err != M_OK)
{
msg = "update file system manager failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
err = GPFSHandler::getPollHandler()->updatePolicyInfo(clusterInfo); // to get policy info
if(err != M_OK)
{
msg = "update policy info failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
err = GPFSHandler::getPollHandler()->updateFilesystemConfigInfo(clusterInfo);// update fs config
if(err != M_OK)
{
msg = "update file system config failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
ClusterStatus* clusterStatus = new ClusterStatus();
err = GPFSHandler::getPollHandler()->getClusterStatus(clusterStatus);
if(err != M_OK)
{
msg = "get cluster status failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
clusterName = clusterInfo->getName();
int i = 0;
string clusterid = clusterInfo->getId();
nFSs = clusterInfo->getNumFilesystems();
//log fs one by one
for( i = 0 ; i < nFSs; i++)
{
fsInfo = clusterInfo->getFilesystem(i);
if (fsInfo == NULL)
{
msg = "NULL filesystem ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&i);
log_error(msg);
continue;
}
fsName = fsInfo->getName();
err = GPFSHandler::getPollHandler()->updateStoragePoolInfo(clusterInfo, (char*)fsName.c_str());
if(err != M_OK)
{
msg = "update storage pool info for file system: ";
msg += fsName;
msg += " failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
continue;
}
msg = "Refresh file system: ";
msg += fsName;
log_debug(msg);
ret = refreshFS(fsInfo, clusterid);
if(ret != TEAL_SUCCESS)
{
msg = "Refresh file system: ";
msg += fsName;
msg += " failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&ret);
log_error(msg);
}
nPools = fsInfo->getNumStoragePools();
int j = 0;
//log stg one by one
for(; j < nPools; j++ )
{
stgInfo = fsInfo->getStoragePool(j);
if(stgInfo == NULL)
{
msg = "ERR stgInfo for storage pool: ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&j);
msg += " in (fs: ";
msg += fsName;
msg += ") is NULL";
log_error(msg);
continue;
}
stgName = stgInfo->getName();
err = GPFSHandler::getPollHandler()->updateDiskInfo(clusterInfo, (char*)fsName.c_str(), (char*)stgName.c_str(),1);
if(err != M_OK)
{
msg = "update disk info in (file system: ";
msg += fsName;
msg += ", storage pool: ";
msg += stgName;
msg += ") failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
continue;
}
msg = "Refresh storage pool: ";
msg += stgName;
msg += " in (fs: ";
msg += fsName;
msg += ")";
log_debug(msg);
ret = refreshStgPool(stgInfo, clusterid, fsName);
if(ret != TEAL_SUCCESS)
{
msg = "Refresh storage pool: ";
msg += stgName;
msg += " in (fs: ";
msg += fsName;
msg += ") failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&ret);
log_error(msg);
}
int k = 0;
nDisks = stgInfo->getNumDisks();
//log disk one by one
for(; k < nDisks ; k++ )
{
diskInfo = stgInfo->getDisk(k);
if(diskInfo == NULL)
{
msg = "diskInfo for disk: ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&k);
msg += " in (storage pool: ";
msg += stgName;
msg += ", fs: ";
msg += fsName;
msg += ") is NULL";
log_error(msg);
continue;
}
diskName = diskInfo->getName();
msg = "Refresh disk: ";
msg += diskName;
msg += " in (storage pool: ";
msg += stgName;
msg += ", fs: ";
msg += fsName;
msg += ")";
log_debug(msg);
ret = refreshDisk(diskInfo, clusterid);
if(ret != TEAL_SUCCESS)
{
msg = "Refresh disk: ";
msg += diskName;
msg += " in (storage pool: ";
msg += stgName;
msg += ", fs: ";
msg += fsName;
msg += ") failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&ret);
log_error(msg);
}
}//end of refresh disks
}//end of refresh stgpool
/* core dump in GPFS 3.4, only effective in 3.5
err = GPFSHandler::getPollHandler()->getFileSets((char*)fsName.c_str(), &fileSetList);
if(err != M_OK)
{
msg = "update fileset info in (fs: ";
msg += fsName;
msg += ") failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_error(msg);
nFsets = 0;
fileSetList = NULL;
continue;
} //at first time to get nFsets but will not return M_OK
*/
err = GPFSHandler::getPollHandler()->getFileSets1((char*)fsName.c_str(), fileSetList, &nFsets);
if(nFsets <= 0)
{
msg = "no fileset found in (fs: ";
msg += fsName;
msg += ")";
log_warn(msg);
nFsets = 0;
fileSetList = NULL;
continue;
}
fileSetList = new FileSet[nFsets];
err = GPFSHandler::getPollHandler()->getFileSets1((char*)fsName.c_str(), fileSetList, &nFsets);
if(err != M_OK)
{
msg = "update fileset info in (fs: ";
msg += fsName;
msg += ") failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
nFsets = 0;
fileSetList = NULL;
continue;
}
int l = 0;
//log fileset one by one
for(; l < nFsets; l++ )
{
fsetInfo = &fileSetList[l];
if(fsetInfo == NULL)
{
msg = "fsetInfo for fset: ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&i);
msg += " in (fs: ";
msg += fsName;
msg += ") is NULL";
log_error(msg);
continue;
}
fsetName = fsetInfo->getName();
msg = "Refresh fileset: ";
msg += fsetName;
msg += " in (fs: ";
msg += fsName;
msg += ")";
log_debug(msg);
ret = refreshFset(fsetInfo, clusterid);
if(ret != TEAL_SUCCESS)
{
msg = "Refresh file set: ";
msg += fsetName;
msg += " in (fs: ";
msg += fsName;
msg += ") failed with";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&ret);
log_error(msg);
}
}//end of refresh fset
if(fileSetList)
{
delete []fileSetList;
fileSetList = NULL;
nFsets = 0;
fsetInfo = NULL;
}
}//end of refresh fs
nNodes = clusterInfo->getNumNodes();
// to get disk access info, place this here to update num_access_disk in nodeinfo and need to invoke updateStoragePool() prior to this API
err = GPFSHandler::getPollHandler()->updateDiskAccessInfo(clusterInfo);
if(err != M_OK)
{
msg = "update disk access info failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_warn(msg);
/* Simply ignore this error to continue....
log_error(msg);
return;
*/
}
NodeInfo* nodeInfo = NULL;
//log node one by one
for( i = 0 ; i < nNodes; i++)
{
nodeInfo = clusterInfo->getNode(i);
if (nodeInfo == NULL)
{
msg = "nodeInfo for node ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&i);
msg += "is NULL";
log_error(msg);
continue;
}
nodeName = nodeInfo->getName();
msg = "Refresh node: ";
msg += nodeName;
log_debug(msg);
ret = refreshNode(nodeInfo, clusterid);
if(ret != TEAL_SUCCESS)
{
msg = "Refresh node: ";
msg += nodeName;
msg += " failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&ret);
log_error(msg);
continue;
}
}//end of refresh node
//refresh free disks here since free disk number/info can only be got after invoking updateDiskInfo() to all fs/stgpool
err = GPFSHandler::getPollHandler()->updateFreeDiskInfo(clusterInfo);
if(err != M_OK)
{
msg = "update free disk info failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
msg += ", ignore it...";
log_warn(msg);
}
nDisks = clusterInfo->getNumFreeDisks();
int k = 0;
for(; k < nDisks ; k++ )
{
diskInfo = clusterInfo->getFreeDisk(k);
if(diskInfo == NULL)
{
msg = "diskInfo for free disk: ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&i);
msg += " is NULL";
log_error(msg);
continue;
}
diskName = diskInfo->getName();
int s;
int nServers = diskInfo->getNumServerItems();
int nBacks = diskInfo->getNumBackupServerItems();
string node_name;
for(s = 0; s < nServers; s++)
{
DiskServerInfo *ds = diskInfo->getServer(s);
node_name += string(ds->getName()) + string(" ");
}
for(s = 0; s < nBacks; s++)
{
DiskServerInfo *ds = diskInfo->getBackupServer(s);
node_name += string(ds->getName()) + string(" ");
}
msg = "Refresh free disk: ";
msg += "(";
msg += diskName;
msg += ")";
log_debug(msg);
char svrList[NAME_STRING_LEN] = {0};
strcpy(svrList,node_name.c_str());
ret = refreshDisk(diskInfo, clusterid, svrList);
if(ret != TEAL_SUCCESS)
{
msg = "Refresh free disk: ";
msg += "(";
msg += diskName;
msg += ") failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&ret);
log_error(msg);
}
}//end of refresh free disks
//refresh cluster here since free disk number/info can only be got after invoking updateDiskInfo() to all fs/stgpool
msg = "Refresh cluster: ";
msg += clusterName;
log_debug(msg);
ret = refreshCluster(clusterInfo,clusterStatus);
if(ret != TEAL_SUCCESS)
{
msg = "Refresh cluster failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&err);
log_error(msg);
}
log_info("##################Start to refresh perseus configuration###################");
int nRgAllocated = 6; /* number of rg slots allocated in the buffer in advance*/
char *bufP = NULL;
int bufLen = 0;
int rc = 0;
int nPdisk = 0;
int nVdisk = 0;
int nRg = 0;
int nDa = 0;
string pdiskName;
string vdiskName;
string rgName;
string daName;
gpfsRecoveryGroupSdrInfo *rgSdrTableP = NULL;
gpfsRecoveryGroupSdrInfo *rgSdrP = NULL;
gpfsRecoveryGroup *rgTableP = NULL;
gpfsRecoveryGroup *rgP = NULL;
gpfsRecoveryGroupDeclusteredArray* daP = NULL;
gpfsDeclusteredArrayPdisk* pdiskP = NULL;
gpfsDeclusteredArrayVdisk* vdiskP = NULL;
rgSdrTableP = new gpfsRecoveryGroupSdrInfo[nRgAllocated];
nRg = nRgAllocated;
/* get initial info from SDR (all RG names) */
rc = getNsdRAIDSdrInfo(rgSdrTableP, &nRg);
// retry if failed with ENOMEM
if(rc == ENOMEM)
{
log_debug("Not enough memory allocated, reallocate...");
nRgAllocated = nRg > nRgAllocated ? nRg : nRgAllocated;
delete[] rgSdrTableP;
rgSdrTableP = NULL;
rgSdrTableP = new gpfsRecoveryGroupSdrInfo[nRgAllocated];
nRg = nRgAllocated;
rc = getNsdRAIDSdrInfo(rgSdrTableP, &nRg);
}
if (rc == M_OK)
{
if (nRg >= 1)
{
rgTableP = new gpfsRecoveryGroup[nRg];
if (rgTableP == NULL)
{
msg = "Initial RG table failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&rc);
log_error(msg);
return;
}
for (i = 0, rgSdrP = rgSdrTableP; i < nRg && i < nRgAllocated; i++, rgSdrP++)
{
rgP = rgTableP + i;
rgP->updateRgSdrInfo(rgSdrP->getRecoveryGroupName(),rgSdrP->getRecoveryGroupServerList(),rgSdrP->getRecoveryGroupId());
rc = getRecoveryGroupSummary(rgP); //refresh rg info
if (rc == 0)
{
rgName = rgP->getRecoveryGroupName();
rc = getRecoveryGroupDeclusteredArrays(rgP); // refresh da info
if (rc == 0)
{
int l = 0;
int nDa = rgP->getRecoveryGroupDeclusterArrays();
bool allDaOK = true; // is all DA ok?
for(; l < nDa; l++)
{
daP = rgP->getDeclusteredArrayP(l);
if(daP == NULL)
{
msg = "da: ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&l);
msg += "in (rg: ";
msg += rgName;
msg += ") is NULL";
log_error(msg);
continue;
}
daName = daP->getDeclusteredArrayName();
msg = "Refresh da: ";
msg += daName;
msg += " in rg: ";
msg += rgName;
log_debug(msg);
ret = refreshDa(daP, clusterid, rgName);
if(ret != TEAL_SUCCESS)
{
msg = "Refresh declustered array: ";
msg += daName;
msg += " in (rg: ";
msg += rgName;
msg += ") failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&ret);
log_error(msg);
}
int j = 0;
int k = 0;
nPdisk = daP->getDeclusteredArrayPdisks();
nVdisk = daP->getDeclusteredArrayVdisks();
for(; j < nPdisk; j++)
{
pdiskP = daP->getDeclusteredArrayPdiskP(j);
if(pdiskP == NULL)
{
msg = "pdisk: ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&j);
msg += " in (rg: ";
msg += rgName;
msg += ", da: ";
msg += daName;
msg += ") is NULL";
log_error(msg);
continue;
}
pdiskName = pdiskP->getPdiskName();
msg = "Refresh pdisk: ";
msg += pdiskName;
msg += " in (rg: ";
msg += rgName;
msg += ", da: ";
msg += daName;
msg += ")";
log_debug(msg);
ret = refreshPdisk(pdiskP,clusterid,rgName,daName);
if(ret != TEAL_SUCCESS)
{
msg = "Refresh pdisk: ";
msg += pdiskName;
msg += " in (rg: ";
msg += rgName;
msg += ", da: ";
msg += daName;
msg += Utils::int_to_char(tmp,10,(unsigned int*)&ret);
log_error(msg);
}
}
for(; k < nVdisk; k++)
{
vdiskP = daP->getDeclusteredArrayVdiskP(k);
if(vdiskP == NULL)
{
msg = "vdisk: ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&k);
msg += " in (rg: ";
msg += rgName;
msg += ", da: ";
msg += daName;
msg += ") is NULL";
log_error(msg);
continue;
}
vdiskName = vdiskP->getVdiskName();
msg = "Refresh vdisk: ";
msg += vdiskName;
msg += " in (rg: ";
msg += rgName;
msg += ", da: ";
msg += daName;
log_debug(msg);
ret = refreshVdisk(vdiskP,clusterid,rgName,daName);
if(ret != TEAL_SUCCESS)
{
msg = "Refresh vdisk: ";
msg += vdiskName;
msg += " in (rg: ";
msg += rgName;
msg += ", da: ";
msg += daName;
msg += ") failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&ret);
log_error(msg);
}
}
allDaOK &= strcmp(daP->getDeclusteredNeedsService(),"yes"); // check all DA's status
}
msg = "Refresh rg: ";
msg += rgName;
log_debug(msg);
ret = refreshRg(rgP, clusterid,allDaOK);
if(ret != TEAL_SUCCESS)
{
msg = "Refresh recovery group: ";
msg += rgName;
msg += " failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&ret);
log_error(msg);
}
}
else
{
msg = "get DA to refresh DA in RG: ";
msg += rgName;
msg += " failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&rc);
log_warn(msg);
continue;
}
}
else
{
msg = "get RG summary to refresh RG: ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&i);
msg += " failed with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&rc);
log_warn(msg);
continue;
}
}
}
else
{
log_warn("No recovery group found!");
}
}
else if(rc == ENODEV)
{
msg = "No perseus configuration..";
log_info(msg);
}
else
{
msg = "Failed to getNsdRAIDSdrInfo with ";
msg += Utils::int_to_char(tmp,10,(unsigned int*)&rc);
log_warn(msg);
}
log_info("########################End of refresh all entities#########################################");
return;
}