You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues-all@impala.apache.org by "ASF subversion and git services (Jira)" <ji...@apache.org> on 2023/01/25 23:50:00 UTC

[jira] [Commented] (IMPALA-11860) Query failure due to large buffer allocation provides insufficient context

    [ https://issues.apache.org/jira/browse/IMPALA-11860?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17680836#comment-17680836 ] 

ASF subversion and git services commented on IMPALA-11860:
----------------------------------------------------------

Commit 37d47d87ff928c21660fadc35a7a506a7bd8d10a in impala's branch refs/heads/master from Michael Smith
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=37d47d87f ]

IMPALA-11860: Add buffer size to thrift serialize error

Add the allocated buffer size to thrift serialization errors to help
debug allocation issues. TMemoryBuffer has a default maxBufferSize of 2
GiB, this can help identify if we're running into issues due to hitting
the max buffer size.

Change-Id: Ie8c1fa7d55197d9e70d21080925238ccdacbeddc
Reviewed-on: http://gerrit.cloudera.org:8080/19440
Reviewed-by: Csaba Ringhofer <cs...@cloudera.com>
Tested-by: Impala Public Jenkins <im...@cloudera.com>


> Query failure due to large buffer allocation provides insufficient context
> --------------------------------------------------------------------------
>
>                 Key: IMPALA-11860
>                 URL: https://issues.apache.org/jira/browse/IMPALA-11860
>             Project: IMPALA
>          Issue Type: Bug
>          Components: Backend
>    Affects Versions: Impala 4.2.0
>            Reporter: Michael Smith
>            Assignee: Michael Smith
>            Priority: Major
>             Fix For: Impala 4.3.0
>
>
> When a query fails because SerializeToBuffer returns a "not OK" status in https://github.com/apache/impala/blob/4.2.0/be/src/runtime/query-state.cc#L646-L655, profile_len is always 0.
> Usually if SerializeToBuffer returns "not OK" it's because it failed to allocate more buffer space, and often has already allocated some buffer. It would help to know how much buffer is allocated, so SerializeToBuffer should update profile_len before returning an error.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-all-unsubscribe@impala.apache.org
For additional commands, e-mail: issues-all-help@impala.apache.org