[
https://issues.apache.org/jira/browse/IMPALA-12056?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Abhishek Rawat updated IMPALA-12056:
------------------------------------
Description:
'Compute Stats' queries gets scheduled on the smallest executor group set since
these queries don't do any real work. However their child queries also gets
scheduled on the smallest executor group. This may not be ideal for cases where
the child query does NDVs and Counts on a big wide table. We should assign the
executor group set for the children queries based on their planning estimates.
Today we see following message for the parent query, which. makes sense:
{code:java}
Verdict: Assign to first group because query is not auto-scalable {code}
as a side effect we see following message for child queries:
{code:java}
Verdict: query option REQUEST_POOL=root.group-set-small is set. Memory and cpu
limit checking is skipped. {code}
was:
'Compute Stats' queries gets scheduled on the smallest executor group set since
these queries don't do any real work. However their child queries also gets
scheduled on the smallest executor group. This may not be ideal for cases where
the child query does an NDV and Count on a big wide table. We should assign the
executor group set for the children queries based on their planning estimates.
Today we see following message for the parent query, which. makes sense:
{code:java}
Verdict: Assign to first group because query is not auto-scalable {code}
as a side effect we see following message for child queries:
{code:java}
Verdict: query option REQUEST_POOL=root.group-set-small is set. Memory and cpu
limit checking is skipped. {code}
> Child queries could get scheduled on improper executor group sets
> -----------------------------------------------------------------
>
> Key: IMPALA-12056
> URL: https://issues.apache.org/jira/browse/IMPALA-12056
> Project: IMPALA
> Issue Type: Bug
> Reporter: Abhishek Rawat
> Priority: Critical
>
> 'Compute Stats' queries gets scheduled on the smallest executor group set
> since these queries don't do any real work. However their child queries also
> gets scheduled on the smallest executor group. This may not be ideal for
> cases where the child query does NDVs and Counts on a big wide table. We
> should assign the executor group set for the children queries based on their
> planning estimates.
> Today we see following message for the parent query, which. makes sense:
> {code:java}
> Verdict: Assign to first group because query is not auto-scalable {code}
> as a side effect we see following message for child queries:
> {code:java}
> Verdict: query option REQUEST_POOL=root.group-set-small is set. Memory and
> cpu limit checking is skipped. {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
