为了更快的定位您的问题,请提供以下信息,谢谢
【详述】执行UPDATE和DELETE都会导致CN节点异常退出
【背景】
【业务影响】
【是否存算分离】是
【StarRocks版本】3.2
【集群规模】例如:1fe(1 follower)+1cn
【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%’;
|enable_pipeline_engine|true|
|—|---|
|max_pipeline_dop|64|
|pipeline_dop|0|
|pipeline_profile_level|1|
|pipeline_sink_dop|0| -
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
UPDATE cq_bigdata.biz_test
SET valueType=‘xxx’
WHERE
bizProcessCode=‘all’
AND tablename=‘all’
and dim=‘all’
ANDcode
=‘divisionCode’
and tenantId=‘1’
- query_dump,怎么获取query_dump文件
-
be crash
- be.out
*** Aborted at 1717733932 (unix time) try “date -d @1717733932” if you are using GNU date ***
PC: @ 0x6b76d5d ra_init
*** SIGILL (@0x6b76d5d) received by PID 175911 (TID 0x14d7f4dc8700) from PID 112684381; stack trace: ***
@ 0x64865a2 google::(anonymous namespace)::FailureSignalHandler()
@ 0x14d917916930 (unknown)
@ 0x6b76d5d ra_init
@ 0x6b76d9b ra_init_with_capacity
@ 0x4da0bad starrocks::DelVector::_add_dels()
@ 0x4da119b starrocks::DelVector::add_dels_as_new_version()
@ 0x50ce68b starrocks::lake::UpdateManager::publish_primary_key_tablet()
@ 0x5108c4e starrocks::lake::PrimaryKeyTxnLogApplier::apply()
@ 0x50bed53 starrocks::lake::publish_version()
@ 0x2d7ceb8 _ZZN9starrocks15LakeServiceImpl15publish_versionEPN6google8protobuf13RpcControllerEPKNS_21PublishVersionRequestEPNS_22PublishVersionResponseEPNS2_7ClosureEENKUlvE_clEv
@ 0x2d7ea9a _ZNSt17_Function_handlerIFvvEZN9starrocks33ConcurrencyLimitedThreadPoolToken11submit_funcESt8functionIS0_ENSt6chrono10time_pointINS5_3_V212system_clockENS5_8durationIlSt5ratioILl1ELl1000000000EEEEEEEUlvE_E9_M_invokeERKSt9_Any_data
@ 0x2e21baf starrocks::ThreadPool::dispatch_thread()
@ 0x2e1b97a starrocks::supervise_thread()
@ 0x14d91790bfed (unknown)
@ 0x14d91769218f clone
@ 0x0 (unknown)
- be.out
-
建表语句
CREATE TABLEcq_bigdata
.biz_test
(bizProcessCode
STRING NOT NULL COMMENT ‘业务过程编码’,tablename
STRING NOT NULL COMMENT ‘指标所在的表名: 如 dws_org_case’,dim
STRING NOT NULL COMMENT ‘维度: 如 org, user, all’,code
STRING NOT NULL COMMENT ‘指标名: 如 inspTaskNum’,tenantId
STRING NOT NULL COMMENT ‘租户ID’,id
BIGINT NOT NULL AUTO_INCREMENT,name
STRING NOT NULL COMMENT ‘指标中文名’,type
STRING NOT NULL COMMENT ‘指标类型: indicator, display’,valueType
STRING NOT NULL COMMENT ‘指标值类型: int, str, dec’,tableAlias
STRING NOT NULL COMMENT ‘表别名: 如 T’,formula
STRING NULL COMMENT ‘指标计算公式’,bizDomainCode
STRING NOT NULL COMMENT ‘所属业务域编码’,dataDomainCode
STRING NOT NULL COMMENT ‘所属数据域编码’,categoryCode
STRING NULL COMMENT ‘指标分类编码’
)
PRIMARY KEY (bizProcessCode
, tablename
, dim
, code
, tenantId
)
COMMENT ‘指标元数据’
/PARTITION BY (tenantId
, dt
)/
DISTRIBUTED BY HASH (code
)
ORDER BY (tenantId
, bizProcessCode
, dim
, code
)
PROPERTIES (
“enable_persistent_index” = “true”
)
;