[
https://issues.apache.org/jira/browse/KUDU-1461?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15296612#comment-15296612
]
Todd Lipcon commented on KUDU-1461:
-----------------------------------
Is it possible this is a result of IMPALA-3373? If you've hit that bug, you'll
have multiple entries for columns, and that confuses the insert/update sink
into trying to set an invalid column. That causes a crash that looks like the
one you pasted.
> kudu::KuduPartialRow::SetInt64 will crash impala in some case.
> ---------------------------------------------------------------
>
> Key: KUDU-1461
> URL: https://issues.apache.org/jira/browse/KUDU-1461
> Project: Kudu
> Issue Type: Bug
> Components: client
> Reporter: zhangsong
>
> Met impala crash recently, from coredump info, found it related to
> kudu-client:
> Current thread is native thread
> siginfo:si_signo=SIGSEGV: si_errno=0, si_code=1 (SEGV_MAPERR),
> si_addr=0x0000000000000000
> ..
> Stack: [0x00007f9fad83e000,0x00007f9fae23f000], sp=0x00007f9fae23cdb0, free
> space=10235k
> Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native
> code)
> C [libkudu_client.so.0+0x8c7cd] kudu::KuduPartialRow::SetInt64(int,
> long)+0x2d
> C [impalad+0xa34f69] impala::KuduTableSink::Send(impala::RuntimeState*,
> impala::RowBatch*, bool)+0x7a9
> C [impalad+0x9fb854] impala::PlanFragmentExecutor::OpenInternal()+0x344
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
