-
Notifications
You must be signed in to change notification settings - Fork 3.7k
[fix](pipeline) premature exit causing core dump during concurrent prepare execution #51492
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
|
Thank you for your contribution to Apache Doris. Please clearly describe your PR:
|
|
run buildall |
|
PR approved by at least one committer and no changes requested. |
|
PR approved by anyone and no changes requested. |
TPC-H: Total hot run time: 34934 ms |
TPC-DS: Total hot run time: 185468 ms |
ClickBench: Total hot run time: 28.83 s |
|
run buildall |
TPC-H: Total hot run time: 33984 ms |
TPC-DS: Total hot run time: 192891 ms |
ClickBench: Total hot run time: 29.09 s |
|
run beut |
|
@yiguolei There are many failing nereids_function_p0 tests in BE UT, but these tests are unrelated to my current changes. Are these test failures expected? |
|
run buildall |
TPC-H: Total hot run time: 33758 ms |
TPC-DS: Total hot run time: 185527 ms |
|
run beut |
BE UT Coverage ReportIncrement line coverage Increment coverage report
|
BE Regression && UT Coverage ReportIncrement line coverage Increment coverage report
|
|
run arm |
|
run performance |
…epare execution (apache#51492) Issue Number: close apache#51491 Problem Summary: When the queue of the FragmentMgrAsync thread pool is full, newly submitted tasks are rejected and return early. However, previously submitted tasks may still be scheduled for execution later. This can lead to premature destruction of objects such as PipelineFragmentContext and TPipelineFragmentParams that are referenced by those tasks, resulting in null pointer exceptions during task execution and ultimately causing a coredump. The pr policy is to wait until all previously submitted tasks are completed before returning. ``` *** SIGSEGV address not mapped to object (@0x1c8) received by PID 3941201 (TID 2115617 OR 0xfe1685bb97f0) from PID 456; stack trace: *** 0# doris::signal::(anonymous namespace)::FailureSignalHandler(int, siginfo_t*, void*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/common/signal_handler.h:421 1# os::Linux::chained_handler(int, siginfo_t*, void*) in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 2# JVM_handle_linux_signal in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 3# signalHandler(int, siginfo_t*, void*) in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 4# 0x0000FFFF6B2A07C0 in linux-vdso.so.1 5# doris::TUniqueId::TUniqueId(doris::TUniqueId const&) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/gensrc/build/gen_cpp/Types_types.cpp:2354 6# doris::AttachTask::AttachTask(doris::QueryContext*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/runtime/thread_context.cpp:60 7# std::_Function_handler<void (), doris::pipeline::PipelineXFragmentContext::_build_pipeline_x_tasks(doris::TPipelineFragmentParams const&, doris::ThreadPool*)::$_0>::_M_invoke(std::_Any_data const&) at /usr/lib/gcc/aarch64-linux-gnu/13/../../../../include/c++/13/bits/std_function.h:290 8# doris::ThreadPool::dispatch_thread() at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/util/threadpool.cpp:552 9# doris::Thread::supervise_thread(void*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/util/thread.cpp:499 10# 0x0000FFFF6AF187AC in /lib64/libpthread.so.0 11# 0x0000FFFF6B16548C in /lib64/libc.so.6 ``` Co-authored-by: XLPE <weiwh1@chinatelecom.cn>
…epare execution (apache#51492) Issue Number: close apache#51491 Problem Summary: When the queue of the FragmentMgrAsync thread pool is full, newly submitted tasks are rejected and return early. However, previously submitted tasks may still be scheduled for execution later. This can lead to premature destruction of objects such as PipelineFragmentContext and TPipelineFragmentParams that are referenced by those tasks, resulting in null pointer exceptions during task execution and ultimately causing a coredump. The pr policy is to wait until all previously submitted tasks are completed before returning. ``` *** SIGSEGV address not mapped to object (@0x1c8) received by PID 3941201 (TID 2115617 OR 0xfe1685bb97f0) from PID 456; stack trace: *** 0# doris::signal::(anonymous namespace)::FailureSignalHandler(int, siginfo_t*, void*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/common/signal_handler.h:421 1# os::Linux::chained_handler(int, siginfo_t*, void*) in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 2# JVM_handle_linux_signal in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 3# signalHandler(int, siginfo_t*, void*) in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 4# 0x0000FFFF6B2A07C0 in linux-vdso.so.1 5# doris::TUniqueId::TUniqueId(doris::TUniqueId const&) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/gensrc/build/gen_cpp/Types_types.cpp:2354 6# doris::AttachTask::AttachTask(doris::QueryContext*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/runtime/thread_context.cpp:60 7# std::_Function_handler<void (), doris::pipeline::PipelineXFragmentContext::_build_pipeline_x_tasks(doris::TPipelineFragmentParams const&, doris::ThreadPool*)::$_0>::_M_invoke(std::_Any_data const&) at /usr/lib/gcc/aarch64-linux-gnu/13/../../../../include/c++/13/bits/std_function.h:290 8# doris::ThreadPool::dispatch_thread() at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/util/threadpool.cpp:552 9# doris::Thread::supervise_thread(void*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/util/thread.cpp:499 10# 0x0000FFFF6AF187AC in /lib64/libpthread.so.0 11# 0x0000FFFF6B16548C in /lib64/libc.so.6 ``` Co-authored-by: XLPE <weiwh1@chinatelecom.cn>
#52365) …epare execution (#51492) Issue Number: close #51491 Problem Summary: When the queue of the FragmentMgrAsync thread pool is full, newly submitted tasks are rejected and return early. However, previously submitted tasks may still be scheduled for execution later. This can lead to premature destruction of objects such as PipelineFragmentContext and TPipelineFragmentParams that are referenced by those tasks, resulting in null pointer exceptions during task execution and ultimately causing a coredump. The pr policy is to wait until all previously submitted tasks are completed before returning. ``` *** SIGSEGV address not mapped to object (@0x1c8) received by PID 3941201 (TID 2115617 OR 0xfe1685bb97f0) from PID 456; stack trace: *** 0# doris::signal::(anonymous namespace)::FailureSignalHandler(int, siginfo_t*, void*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/common/signal_handler.h:421 1# os::Linux::chained_handler(int, siginfo_t*, void*) in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 2# JVM_handle_linux_signal in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 3# signalHandler(int, siginfo_t*, void*) in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 4# 0x0000FFFF6B2A07C0 in linux-vdso.so.1 5# doris::TUniqueId::TUniqueId(doris::TUniqueId const&) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/gensrc/build/gen_cpp/Types_types.cpp:2354 6# doris::AttachTask::AttachTask(doris::QueryContext*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/runtime/thread_context.cpp:60 7# std::_Function_handler<void (), doris::pipeline::PipelineXFragmentContext::_build_pipeline_x_tasks(doris::TPipelineFragmentParams const&, doris::ThreadPool*)::$_0>::_M_invoke(std::_Any_data const&) at /usr/lib/gcc/aarch64-linux-gnu/13/../../../../include/c++/13/bits/std_function.h:290 8# doris::ThreadPool::dispatch_thread() at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/util/threadpool.cpp:552 9# doris::Thread::supervise_thread(void*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/util/thread.cpp:499 10# 0x0000FFFF6AF187AC in /lib64/libpthread.so.0 11# 0x0000FFFF6B16548C in /lib64/libc.so.6 ``` Co-authored-by: XLPE <crykix@gmail.com> Co-authored-by: XLPE <weiwh1@chinatelecom.cn>
apache#52365) …epare execution (apache#51492) Issue Number: close apache#51491 Problem Summary: When the queue of the FragmentMgrAsync thread pool is full, newly submitted tasks are rejected and return early. However, previously submitted tasks may still be scheduled for execution later. This can lead to premature destruction of objects such as PipelineFragmentContext and TPipelineFragmentParams that are referenced by those tasks, resulting in null pointer exceptions during task execution and ultimately causing a coredump. The pr policy is to wait until all previously submitted tasks are completed before returning. ``` *** SIGSEGV address not mapped to object (@0x1c8) received by PID 3941201 (TID 2115617 OR 0xfe1685bb97f0) from PID 456; stack trace: *** 0# doris::signal::(anonymous namespace)::FailureSignalHandler(int, siginfo_t*, void*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/common/signal_handler.h:421 1# os::Linux::chained_handler(int, siginfo_t*, void*) in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 2# JVM_handle_linux_signal in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 3# signalHandler(int, siginfo_t*, void*) in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 4# 0x0000FFFF6B2A07C0 in linux-vdso.so.1 5# doris::TUniqueId::TUniqueId(doris::TUniqueId const&) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/gensrc/build/gen_cpp/Types_types.cpp:2354 6# doris::AttachTask::AttachTask(doris::QueryContext*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/runtime/thread_context.cpp:60 7# std::_Function_handler<void (), doris::pipeline::PipelineXFragmentContext::_build_pipeline_x_tasks(doris::TPipelineFragmentParams const&, doris::ThreadPool*)::$_0>::_M_invoke(std::_Any_data const&) at /usr/lib/gcc/aarch64-linux-gnu/13/../../../../include/c++/13/bits/std_function.h:290 8# doris::ThreadPool::dispatch_thread() at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/util/threadpool.cpp:552 9# doris::Thread::supervise_thread(void*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/util/thread.cpp:499 10# 0x0000FFFF6AF187AC in /lib64/libpthread.so.0 11# 0x0000FFFF6B16548C in /lib64/libc.so.6 ``` Co-authored-by: XLPE <crykix@gmail.com> Co-authored-by: XLPE <weiwh1@chinatelecom.cn>
…epare execution (apache#51492) Issue Number: close apache#51491 Problem Summary: When the queue of the FragmentMgrAsync thread pool is full, newly submitted tasks are rejected and return early. However, previously submitted tasks may still be scheduled for execution later. This can lead to premature destruction of objects such as PipelineFragmentContext and TPipelineFragmentParams that are referenced by those tasks, resulting in null pointer exceptions during task execution and ultimately causing a coredump. The pr policy is to wait until all previously submitted tasks are completed before returning. ``` *** SIGSEGV address not mapped to object (@0x1c8) received by PID 3941201 (TID 2115617 OR 0xfe1685bb97f0) from PID 456; stack trace: *** 0# doris::signal::(anonymous namespace)::FailureSignalHandler(int, siginfo_t*, void*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/common/signal_handler.h:421 1# os::Linux::chained_handler(int, siginfo_t*, void*) in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 2# JVM_handle_linux_signal in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 3# signalHandler(int, siginfo_t*, void*) in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 4# 0x0000FFFF6B2A07C0 in linux-vdso.so.1 5# doris::TUniqueId::TUniqueId(doris::TUniqueId const&) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/gensrc/build/gen_cpp/Types_types.cpp:2354 6# doris::AttachTask::AttachTask(doris::QueryContext*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/runtime/thread_context.cpp:60 7# std::_Function_handler<void (), doris::pipeline::PipelineXFragmentContext::_build_pipeline_x_tasks(doris::TPipelineFragmentParams const&, doris::ThreadPool*)::$_0>::_M_invoke(std::_Any_data const&) at /usr/lib/gcc/aarch64-linux-gnu/13/../../../../include/c++/13/bits/std_function.h:290 8# doris::ThreadPool::dispatch_thread() at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/util/threadpool.cpp:552 9# doris::Thread::supervise_thread(void*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/util/thread.cpp:499 10# 0x0000FFFF6AF187AC in /lib64/libpthread.so.0 11# 0x0000FFFF6B16548C in /lib64/libc.so.6 ``` Co-authored-by: XLPE <weiwh1@chinatelecom.cn>
#52850) …epare execution (#51492) Issue Number: close #51491 Problem Summary: When the queue of the FragmentMgrAsync thread pool is full, newly submitted tasks are rejected and return early. However, previously submitted tasks may still be scheduled for execution later. This can lead to premature destruction of objects such as PipelineFragmentContext and TPipelineFragmentParams that are referenced by those tasks, resulting in null pointer exceptions during task execution and ultimately causing a coredump. The pr policy is to wait until all previously submitted tasks are completed before returning. ``` *** SIGSEGV address not mapped to object (@0x1c8) received by PID 3941201 (TID 2115617 OR 0xfe1685bb97f0) from PID 456; stack trace: *** 0# doris::signal::(anonymous namespace)::FailureSignalHandler(int, siginfo_t*, void*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/common/signal_handler.h:421 1# os::Linux::chained_handler(int, siginfo_t*, void*) in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 2# JVM_handle_linux_signal in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 3# signalHandler(int, siginfo_t*, void*) in /usr/jdk64/current/jre/lib/aarch64/server/libjvm.so 4# 0x0000FFFF6B2A07C0 in linux-vdso.so.1 5# doris::TUniqueId::TUniqueId(doris::TUniqueId const&) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/gensrc/build/gen_cpp/Types_types.cpp:2354 6# doris::AttachTask::AttachTask(doris::QueryContext*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/runtime/thread_context.cpp:60 7# std::_Function_handler<void (), doris::pipeline::PipelineXFragmentContext::_build_pipeline_x_tasks(doris::TPipelineFragmentParams const&, doris::ThreadPool*)::$_0>::_M_invoke(std::_Any_data const&) at /usr/lib/gcc/aarch64-linux-gnu/13/../../../../include/c++/13/bits/std_function.h:290 8# doris::ThreadPool::dispatch_thread() at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/util/threadpool.cpp:552 9# doris::Thread::supervise_thread(void*) at /home/jenkins_agent/workspace/BigDataComponent_doris-unified-arm-release/be/src/util/thread.cpp:499 10# 0x0000FFFF6AF187AC in /lib64/libpthread.so.0 11# 0x0000FFFF6B16548C in /lib64/libc.so.6 ``` Co-authored-by: XLPE <crykix@gmail.com> Co-authored-by: XLPE <weiwh1@chinatelecom.cn>
What problem does this PR solve?
Issue Number: close #51491
Problem Summary:
When the queue of the FragmentMgrAsync thread pool is full, newly submitted tasks are rejected and return early. However, previously submitted tasks may still be scheduled for execution later. This can lead to premature destruction of objects such as PipelineFragmentContext and TPipelineFragmentParams that are referenced by those tasks, resulting in null pointer exceptions during task execution and ultimately causing a coredump.
The pr policy is to wait until all previously submitted tasks are completed before returning.
Release note
None
Check List (For Author)
Test
Behavior changed:
Does this need documentation?
Check List (For Reviewer who merge this PR)