3.1.11 执行stream load时be崩溃

[daducha]

为了更快的定位您的问题，请提供以下信息，谢谢
【详述】问题详细描述
【背景】做过哪些操作？
【业务影响】
【是否存算分离】
【StarRocks版本】例如：3.1.11
【集群规模】例如：3fe（1 follower+2observer）+5be（fe与be混部）
【机器信息】CPU虚拟核/内存/网卡，例如：48C/64G/万兆
【联系方式】为了在解决问题过程中能及时联系到您获取一些日志信息，请补充下您的联系方式，例如：社区群4-小李或者邮箱，谢谢
【附件】

• fe.log/beINFO/相应截图
• 慢查询：
  • Profile信息，获取Profile，通过Profile分析查询瓶颈
  • 并行度：show variables like ‘%parallel_fragment_exec_instance_num%’;
  • pipeline是否开启：show variables like ‘%pipeline%’;
  • be节点cpu和内存使用率截图
• 查询报错：
  • query_dump，怎么获取query_dump文件
• be crash
  • be.out
  • coredump，如何获取coredump
• 外表查询报错
  • be.out和fe.warn.log

[daducha]

PC: @ 0x3311bb8 starrocks::stream_load::OlapTableSink::_send_chunk_by_node()
*** SIGSEGV (@0x10) received by PID 1025943 (TID 0x7efde6b12700) from PID 16; stack trace: ***
@ 0x651fba2 google::(anonymous namespace)::FailureSignalHandler()
@ 0x7efe9457c370 (unknown)
@ 0x3311bb8 starrocks::stream_load::OlapTableSink::_send_chunk_by_node()
@ 0x33125c7 starrocks::stream_load::OlapTableSink::_send_chunk()
@ 0x3313885 starrocks::stream_load::OlapTableSink::send_chunk()
@ 0x2bf63f2 starrocks::PlanFragmentExecutor::_open_internal_vectorized()
@ 0x2bf8e0d starrocks::PlanFragmentExecutor::open()
@ 0x2b4a31b starrocks::FragmentExecState::execute()
@ 0x2b50ba4 starrocks::FragmentMgr::exec_actual()
@ 0x2d0234c starrocks::ThreadPool::dispatch_thread()
@ 0x2cfbffa starrocks::supervise_thread()
@ 0x7efe94574dc5 start_thread
@ 0x7efe942a373d __clone
@ 0x0 (unknown)
start time: Wed Nov 13 23:14:55 CST 2024

[daducha]

Program terminated with signal SIGSEGV, Segmentation fault.
#0 0x0000000003311bb8 in starrocks::Chunk::num_rows (this=) at /root/starrocks/be/src/column/chunk.h:91
91 /root/starrocks/be/src/column/chunk.h: 没有那个文件或目录.
[Current thread is 1 (Thread 0x7efd223cd700 (LWP 228422))]
(gdb) bt
#0 0x0000000003311bb8 in starrocks::Chunk::num_rows (this=) at /root/starrocks/be/src/column/chunk.h:91
#1 starrocks::stream_load::NodeChannel::_send_request (this=0x7efd58ae2800, eos=141, finished=152) at /root/starrocks/be/src/exec/tablet_sink.cpp:681
#2 0x0000000000000028 in ?? ()
#3 0x00007efd223c6c78 in ?? ()
#4 0x00007efbd83168d0 in ?? ()
#5 0x00007efd58ae29e0 in ?? ()
#6 0x00007efd58ae2a10 in ?? ()
#7 0x00007efda24ed5c0 in ?? ()
#8 0x00007efd988dc800 in ?? ()
#9 0x00007ef9a5c5f140 in ?? ()
#10 0x00007efd9df536c8 in ?? ()
#11 0x0000000000000000 in ?? ()
(gdb) f 1
#1 starrocks::stream_load::NodeChannel::_send_request (this=0x7efd58ae2800, eos=141, finished=152) at /root/starrocks/be/src/exec/tablet_sink.cpp:681
681 /root/starrocks/be/src/exec/tablet_sink.cpp: 没有那个文件或目录.
(gdb) p chunk
$1 = {_M_t = {<std::__uniq_ptr_impl<starrocks::Chunk, std::default_deletestarrocks::Chunk >> = {
_M_t = {<std::_Tuple_impl<0, starrocks::Chunk*, std::default_deletestarrocks::Chunk >> = {<std::_Tuple_impl<1, std::default_deletestarrocks::Chunk >> = {<std::_Head_base<1, std::default_deletestarrocks::Chunk, true>> = {<std::default_deletestarrocks::Chunk> = {}, }, }, <std::_Head_base<0, starrocks::Chunk*, false>> = {
_M_head_impl = 0x0}, }, }}, }}
gdb core文件发现chunk是空指针

[daducha]

CREATE TABLE tbl (
c1 datetime NOT NULL COMMENT “”,
c2 datetime NOT NULL COMMENT “”,
c3 varchar(65533) NULL COMMENT “”,
c4 varchar(65533) NULL COMMENT “”,
c5 varchar(65533) NULL COMMENT “”,
c6 varchar(65533) NULL COMMENT “”,
c7 varchar(65533) NULL COMMENT “”,
c8 varchar(65533) NULL COMMENT “”,
c9 varchar(65533) NULL COMMENT “”,
c10 varchar(65533) NULL COMMENT “”,
c11 varchar(65533) NULL COMMENT “”,
c12 varchar(65533) NULL COMMENT “”,
c13 varchar(65533) NULL COMMENT “”,
c14 varchar(65533) NULL COMMENT “”,
c15 varchar(65533) NULL COMMENT “”,
c16 varchar(65533) NULL COMMENT “”,
c17 varchar(65533) NULL COMMENT “”,
c18 hll HLL_UNION NULL COMMENT “”,
c19 hll HLL_UNION NULL COMMENT “”,
c20 hll HLL_UNION NULL COMMENT “”,
c21 hll HLL_UNION NULL COMMENT “”,
c22 hll HLL_UNION NULL COMMENT “”,
c23 hll HLL_UNION NULL COMMENT “”
) ENGINE=OLAP
AGGREGATE KEY(c1, c2, c3, c4, c5, c6, c7, c8, c9, c10, c11, c12, c13, c14, c15, c16, c17)
COMMENT “hdp_lbg_ectech_lego_ods_req_dsp_new”
PARTITION BY date_trunc(‘hour’, c1)
DISTRIBUTED BY HASH(c2, c3, c4) BUCKETS 10
PROPERTIES (
“replication_num” = “3”,
“in_memory” = “false”,
“enable_persistent_index” = “false”,
“replicated_storage” = “true”,
“partition_live_number” = “336”,
“compression” = “LZ4”
);
表结构

[daducha]

@jingdan 大佬帮忙看看是不是解决过