3014524 - How-To - Analyzing the 'INDEXMANAGER_WAITGRAPH' contents in HANA Runtime Dumps

• You observe a long lock-wait situation within the '[INDEXMANAGER_WAITGRAPH]' section of a classic HANA runtime dump or the specific runtime dump trace for blocked index handles :

[INDEXMANAGER_WAITGRAPH] Wait graph for index handles: (2020-12-15 03:01:29 652 Local)
 digraph WaitGraph {
 n0x00007daeded15000[shape=record,label="{SqlExecutor\nSAPHANADB:KAPA:search_delta\nsyscall\nSynchronization::BinarySemaphore::timedWait\nUnifiedTable::PageChainBase\<UnifiedTable::PageDirectoryElement\>::loadTemp\nUnifiedTable::SerialDataStorage::readNextBufferImpl\nAttributeEngine::FileReadBuffer::readBlock\nAttributeEngine::SpAttribute\<TrexTypes::StringAttributeValue, AttributeEngine::ValueDict\<TrexTypes::StringAttributeValue\>, 
AttributeEngine::RleMemoryDocuments\>::load\nAttributeEngine::MemoryAvc2::load_internal\nAttributeEngine::MemoryAvc2::load\nAttributeEngine::MemoryAvc2::open\nAttributeEngine::ColumnResource::loadAVC\nAttributeEngine::ColumnList::getAVCImpl\nAttributeEngine::AttributeApi::openAttributeAndUpdateReleaseTime\nAttributeEngine::AttributeApi::sortDocuments\
nqo3::AttributeApiWrapper::sortDocuments\nqo3::OptimizerImpl::doSort\nqo3::OptimizerImpl::evaluate\nqo3::Optimizer::evaluate\nqo3::Optimizer::evaluate\nTRexAPI::Search::OptimizerImpl::evaluate\nTRexAPI::\nTRexAPI::LocalSearchImpl::doSearch\nTRexAPI::SearchAPI::singlePartSearch\
nTRexAPI::SearchAPI::extractResults\nTRexAPI::SearchAPI::open\nptime::TrexOltpSearch::open\nptime::Trex_oltp_search::do_open\nptime::Table::open\nptime::Query::result_t::open\nptime::ResultSet::open_\nptime::Statement::post_execute_\nptime::Statement::execute_\nptime::PreparedStatement::execute\nptime::ActionExecutor::executeQuery\nptime::ActionExecutor::executeQid\nptime::ActionExecutor::executeAction\nptime::SessionHandler::executeAction\
n.LTHUNK390.lto_priv.854\nptime::SessionHandler::receiveMessage_\nptime::SessionHandler::receiveMessage\nptime::TcpReceiver::doWork\nptime::TcpReceiver::run\nTrexThreads::PoolThread::run\
nTrexThreads::PoolThread::run\nExecution::Thread::staticMainImp\nExecution::pthreadFunctionWrapper\nstart_thread\n__clone\n}",style=filled,fillcolor="#ffffff",fontname=sans]

...... 

n0x00007f6797b2b000 -> n0x00007ef6ddd9a000 [label="waits for TRexConfig_IndexMgrIndex_ReplayLock\n(owner locked exclusively)"]
 n0x00007f7552820000 -> n0x00007f73e648a000 [label="waits for LoadLock\n(owner locked exclusively)"]
 n0x00007f84965ce000 -> n0x00007e47f939f000 [label="waits for LoadLock\n(owner locked exclusively)"]
 }
 [OK]

• The format of the file is identified by this trace file name "<service>_<host>.<port>.rtedump.<timestamp>.trc" or "<service>_<host>.<port>.rtedump.<timestamp>.indexhandle_blocked.trc"

Image/data in this KBA is from SAP internal systems, sample data, or demo systems. Any resemblance to real data is purely coincidental. 

• SAP HANA PLATFORM EDITION 1.0
• SAP HANA PLATFORM EDITION 2.0

lock wait, system hang, system slow, long running statement, hanging, transaction hanging, dump, analyzer, waitgraph, wait, graph, longrunning , KBA , HAN-DB-ENG , SAP HANA DB Engines , HAN-DB-PERF , SAP HANA Database Performance , Problem