[ https://issues.apache.org/jira/browse/KUDU-2727?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17125466#comment-17125466 ]
Alexey Serbin commented on KUDU-2727: ------------------------------------- Another set of stacks, just for more context (captured with code close to kudu 1.10.1): {noformat} tids=[1866940] 0x7fc8d67f95e0 <unknown> 0x1ec35f4 base::internal::SpinLockDelay() 0x1ec347c base::SpinLock::SlowLock() 0xb68518 kudu::consensus::RaftConsensus::NotifyCommitIndex() 0xb4c9e7 kudu::consensus::PeerMessageQueue::NotifyObserversTask() 0xb47850 _ZN4kudu8internal7InvokerILi2ENS0_9BindStateINS0_15RunnableAdapterIMNS_9consensus16PeerMessageQueueEFvRKSt8functionIFvPNS4_24PeerMessageQueueObserverEEEEEEFvPS5_SC_EFvNS0_17UnretainedWrapperIS5_EEZNS5_34NotifyObserversOfCommitIndexChangeElEUlS8_E_EEESH_E3RunEPNS0_13BindStateBaseE 0x1eaedbf kudu::ThreadPool::DispatchThread() 0x1ea4a84 kudu::Thread::SuperviseThread() 0x7fc8d67f1e25 start_thread 0x7fc8d4acf34d __clone tids=[1370336,1370326,1370327,1370328,1370329,1370330,1370331,1370332,1370333,1370334,1370335,1370323,1370337,1370338,1370339,1370340,1370341,1370342,1370343,1370344,1370345,1370346,1370353,1370361,1370360,1370359,1370358,1370357,1370356,1370355,1370354,1370325,1370352,1370351,1370350,1370349,1370348,1370347,1370322,1370324] 0x7fc8d67f95e0 <unknown> 0x1ec35f4 base::internal::SpinLockDelay() 0x1ec347c base::SpinLock::SlowLock() 0xb7deb8 kudu::consensus::RaftConsensus::CheckLeadershipAndBindTerm() 0xaab010 kudu::tablet::TransactionDriver::ExecuteAsync() 0xaa344c kudu::tablet::TabletReplica::SubmitWrite() 0x928fb0 kudu::tserver::TabletServiceImpl::Write() 0x1d2e8d9 kudu::rpc::GeneratedServiceIf::Handle() 0x1d2efd9 kudu::rpc::ServicePool::RunThread() 0x1ea4a84 kudu::Thread::SuperviseThread() 0x7fc8d67f1e25 start_thread 0x7fc8d4acf34d __clone tids=[1866932,1866929] 0x7fc8d67f95e0 <unknown> 0x7fc8d67f5943 __pthread_cond_wait 0xb99b38 kudu::log::Log::AsyncAppend() 0xb9c24c kudu::log::Log::AsyncAppendCommit() 0xaad489 kudu::tablet::TransactionDriver::ApplyTask() 0x1eaedbf kudu::ThreadPool::DispatchThread() 0x1ea4a84 kudu::Thread::SuperviseThread() 0x7fc8d67f1e25 start_thread 0x7fc8d4acf34d __clone tids=[1866928] 0x7fc8d67f95e0 <unknown> 0x7fc8d67f5943 __pthread_cond_wait 0xb99b38 kudu::log::Log::AsyncAppend() 0xb9c493 kudu::log::Log::AsyncAppendReplicates() 0xb597e9 kudu::consensus::LogCache::AppendOperations() 0xb4fa24 kudu::consensus::PeerMessageQueue::AppendOperations() 0xb4fd45 kudu::consensus::PeerMessageQueue::AppendOperation() 0xb6f28c kudu::consensus::RaftConsensus::AppendNewRoundToQueueUnlocked() 0xb7dff8 kudu::consensus::RaftConsensus::Replicate() 0xaab8e7 kudu::tablet::TransactionDriver::Prepare() 0xaac009 kudu::tablet::TransactionDriver::PrepareTask() 0x1eaedbf kudu::ThreadPool::DispatchThread() 0x1ea4a84 kudu::Thread::SuperviseThread() {noformat} > Contention on the Raft consensus lock can cause tablet service queue overflows > ------------------------------------------------------------------------------ > > Key: KUDU-2727 > URL: https://issues.apache.org/jira/browse/KUDU-2727 > Project: Kudu > Issue Type: Improvement > Components: perf > Reporter: William Berkeley > Assignee: Mike Percy > Priority: Major > > Here's stacks illustrating the phenomenon: > {noformat} > tids=[2201] > 0x379ba0f710 <unknown> > 0x1fb951a base::internal::SpinLockDelay() > 0x1fb93b7 base::SpinLock::SlowLock() > 0xb4e68e kudu::consensus::Peer::SignalRequest() > 0xb9c0df kudu::consensus::PeerManager::SignalRequest() > 0xb8c178 kudu::consensus::RaftConsensus::Replicate() > 0xaab816 kudu::tablet::TransactionDriver::Prepare() > 0xaac0ed kudu::tablet::TransactionDriver::PrepareTask() > 0x1fa37ed kudu::ThreadPool::DispatchThread() > 0x1f9c2a1 kudu::Thread::SuperviseThread() > 0x379ba079d1 start_thread > 0x379b6e88fd clone > tids=[4515] > 0x379ba0f710 <unknown> > 0x1fb951a base::internal::SpinLockDelay() > 0x1fb93b7 base::SpinLock::SlowLock() > 0xb74c60 kudu::consensus::RaftConsensus::NotifyCommitIndex() > 0xb59307 kudu::consensus::PeerMessageQueue::NotifyObserversTask() > 0xb54058 > _ZN4kudu8internal7InvokerILi2ENS0_9BindStateINS0_15RunnableAdapterIMNS_9consensus16PeerMessageQueueEFvRKSt8functionIFvPNS4_24PeerMessageQueueObserverEEEEEEFvPS5_SC_EFvNS0_17UnretainedWrapperIS5_EEZNS5_34NotifyObserversOfCommitIndexChangeElEUlS8_E_EEESH_E3RunEPNS0_13BindStateBaseE > 0x1fa37ed kudu::ThreadPool::DispatchThread() > 0x1f9c2a1 kudu::Thread::SuperviseThread() > 0x379ba079d1 start_thread > 0x379b6e88fd clone > tids=[22185,22194,22193,22188,22187,22186] > 0x379ba0f710 <unknown> > 0x1fb951a base::internal::SpinLockDelay() > 0x1fb93b7 base::SpinLock::SlowLock() > 0xb8bff8 > kudu::consensus::RaftConsensus::CheckLeadershipAndBindTerm() > 0xaaaef9 kudu::tablet::TransactionDriver::ExecuteAsync() > 0xaa3742 kudu::tablet::TabletReplica::SubmitWrite() > 0x92812d kudu::tserver::TabletServiceImpl::Write() > 0x1e28f3c kudu::rpc::GeneratedServiceIf::Handle() > 0x1e2986a kudu::rpc::ServicePool::RunThread() > 0x1f9c2a1 kudu::Thread::SuperviseThread() > 0x379ba079d1 start_thread > 0x379b6e88fd clone > tids=[22192,22191] > 0x379ba0f710 <unknown> > 0x1fb951a base::internal::SpinLockDelay() > 0x1fb93b7 base::SpinLock::SlowLock() > 0x1e13dec kudu::rpc::ResultTracker::TrackRpc() > 0x1e28ef5 kudu::rpc::GeneratedServiceIf::Handle() > 0x1e2986a kudu::rpc::ServicePool::RunThread() > 0x1f9c2a1 kudu::Thread::SuperviseThread() > 0x379ba079d1 start_thread > 0x379b6e88fd clone > tids=[4426] > 0x379ba0f710 <unknown> > 0x206d3d0 <unknown> > 0x212fd25 google::protobuf::Message::SpaceUsedLong() > 0x211dee4 > google::protobuf::internal::GeneratedMessageReflection::SpaceUsedLong() > 0xb6658e kudu::consensus::LogCache::AppendOperations() > 0xb5c539 kudu::consensus::PeerMessageQueue::AppendOperations() > 0xb5c7c7 kudu::consensus::PeerMessageQueue::AppendOperation() > 0xb7c675 > kudu::consensus::RaftConsensus::AppendNewRoundToQueueUnlocked() > 0xb8c147 kudu::consensus::RaftConsensus::Replicate() > 0xaab816 kudu::tablet::TransactionDriver::Prepare() > 0xaac0ed kudu::tablet::TransactionDriver::PrepareTask() > 0x1fa37ed kudu::ThreadPool::DispatchThread() > 0x1f9c2a1 kudu::Thread::SuperviseThread() > 0x379ba079d1 start_thread > 0x379b6e88fd clone > {noformat} > {{kudu::consensus::RaftConsensus::CheckLeadershipAndBindTerm()}} needs to > take the lock to check the term and the Raft role. When many RPCs come in for > the same tablet, the contention can hog service threads and cause queue > overflows on busy systems. > Yugabyte switched their equivalent lock to be an atomic that allows them to > read the term and role wait-free. -- This message was sent by Atlassian Jira (v8.3.4#803005)