[
https://issues.apache.org/jira/browse/IMPALA-8294?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16832923#comment-16832923
]
Tim Armstrong commented on IMPALA-8294:
---------------------------------------
If you look in HdfsScanNodeBase, bytes_read_ follows the good pattern where it
gets updated by DiskIoMgr in real time. bytes_read_local_ follows the bad
pattern.
> Inconsistent updating of BytesRead* counters
> --------------------------------------------
>
> Key: IMPALA-8294
> URL: https://issues.apache.org/jira/browse/IMPALA-8294
> Project: IMPALA
> Issue Type: Bug
> Components: Backend
> Affects Versions: Impala 3.0, Impala 2.12.0, Impala 3.1.0, Impala 3.2.0
> Reporter: Lars Volker
> Assignee: Abhishek Rawat
> Priority: Major
> Labels: observability, profile
>
> Some of the BytesRead* counters in profiles (e.g. BytesReadLocal) are only
> updated once a query finishes successfully. This leads to confusion because
> queries that are still running or failed look like they did not read data
> locally.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-all-unsubscr...@impala.apache.org
For additional commands, e-mail: issues-all-h...@impala.apache.org
