CREATE CATALOG 导致be 挂掉,是什么原因?

Viewed 32

CREATE CATALOG mall_product_dev PROPERTIES (
"type"="jdbc",
"user"="root",
"password"="123456",
"jdbc_url" = "jdbc:mysql://192.168.1.7:3306/product?useUnicode=true&characterEncoding=utf8&useSSL=false",
"driver_url" = "https://doris-community-test-1308700295.cos.ap-hongkong.myqcloud.com/jdbc_driver/mysql-connector-java-8.0.25.jar",
"driver_class" = "com.mysql.cj.jdbc.Driver"
);
执行语句 导致be节点挂掉, 版本:doris-2.1.4, 系统centos-7.9, 日志如下:

StdoutLogger 2024-09-27 17:20:02,800 Start time: Fri Sep 27 17:20:02 CST 2024
INFO: java_cmd /usr/local/jdk1.8.0_381/bin/java
INFO: jdk_version 8
SLF4J: Class path contains multiple SLF4J bindings.
SLF4J: Found binding in [jar:file:/usr/local/doris-2.1.4/be/lib/java_extensions/preload-extensions/preload-extensions-jar-with-dependencies.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: Found binding in [jar:file:/usr/local/doris-2.1.4/be/lib/java_extensions/java-udf/java-udf-jar-with-dependencies.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: Found binding in [jar:file:/usr/local/doris-2.1.4/be/lib/hadoop_hdfs/common/lib/slf4j-reload4j-1.7.36.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation.
SLF4J: Actual binding is of type [org.slf4j.impl.Reload4jLoggerFactory]
[WARNING!] /sys/kernel/mm/transparent_hugepage/enabled: [always] madvise never, Doris not recommend turning on THP, which may cause the BE process to use more memory and cannot be freed in time. Turn off THP: echo madvise | sudo tee /sys/kernel/mm/transparent_hugepage/enabled
*** Query id: c67450f87114818-b1344657880e05f7 ***
*** is nereids: 0 ***
*** tablet id: 0 ***
*** Aborted at 1727429105 (unix time) try "date -d @1727429105" if you are using GNU date ***
*** Current BE git commitID: e93678fd1e ***
*** SIGSEGV unknown detail explain (@0x0) received by PID 2147 (TID 4450 OR 0x7efd8b2c7700) from PID 0; stack trace: ***
0# doris::signal::(anonymous namespace)::FailureSignalHandler(int, siginfo_t*, void*) at /home/zcp/repo_center/doris_release/doris/be/src/common/signal_handler.h:421
1# os::Linux::chained_handler(int, siginfo*, void*) in /usr/local/jdk1.8.0_381/jre/lib/amd64/server/libjvm.so
2# JVM_handle_linux_signal in /usr/local/jdk1.8.0_381/jre/lib/amd64/server/libjvm.so
3# signalHandler(int, siginfo*, void*) in /usr/local/jdk1.8.0_381/jre/lib/amd64/server/libjvm.so
4# 0x00007F011B5D8510 in /lib64/libc.so.6
5# oopDesc* JNIHandles::resolve_jweak(jobject*) in /usr/local/jdk1.8.0_381/jre/lib/amd64/server/libjvm.so
6# jni_invoke_nonstatic(JNIEnv
, JavaValue, _jobject*, JNICallType, jmethodID*, JNI_ArgumentPusher*, Thread*) [clone .constprop.108] in /usr/local/jdk1.8.0_381/jre/lib/amd64/server/libjvm.so
7# jni_CallNonvirtualVoidMethodV in /usr/local/jdk1.8.0_381/jre/lib/amd64/server/libjvm.so
8# JNIEnv
::CallNonvirtualVoidMethod(_jobject*, _jclass*, _jmethodID*, ...) at /usr/lib/jvm/java-8-openjdk-amd64/include/jni.h:1265
9# doris::vectorized::JdbcConnector::clean_datasource() at /home/zcp/repo_center/doris_release/doris/be/src/vec/exec/vjdbc_connector.cpp:200
10# std::_Function_handler<void (), doris::PInternalServiceImpl::test_jdbc_connection(google::protobuf::RpcController*, doris::PJdbcTestConnectionRequest const*, doris::PJdbcTestConnectionResult*, google::protobuf::Closure*)::$_0>::_M_invoke(std::_Any_data const&) at /var/local/ldb_toolchain/bin/../lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/bits/std_function.h:291
11# doris::WorkThreadPool::work_thread(int) at /home/zcp/repo_center/doris_release/doris/be/src/util/work_thread_pool.hpp:159
12# execute_native_thread_routine at ../../../../../libstdc++-v3/src/c++11/thread.cc:84
13# start_thread at /usr/local/glibc-2.28/nptl/pthread_create.c:487
14# clone at ../sysdeps/unix/sysv/linux/x86_64/clone.S:97

2 Answers

WARNING!] /sys/kernel/mm/transparent_hugepage/enabled: [always] madvise never, Doris not recommend turning on THP, which may cause the BE process to use more memory and cannot be freed in time. Turn off THP: echo madvise | sudo tee /sys/kernel/mm/transparent_hugepage/enabled

echo never > /sys/kernel/mm/transparent_hugepage/enabled
echo never > /sys/kernel/mm/transparent_hugepage/defrag

把大页禁用再试试呢?