[jira] [Commented] (IMPALA-10504) Add tracing for remote block reads

2021-02-24 Thread ASF subversion and git services (Jira)


[ 
https://issues.apache.org/jira/browse/IMPALA-10504?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17290602#comment-17290602
 ] 

ASF subversion and git services commented on IMPALA-10504:
--

Commit c8d94e6e365ca21d731617767bd7793f67f3 in impala's branch 
refs/heads/master from Kurt Deschler
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=c8d94e6 ]

IMPALA-10504: Add tracing for remote block reads

This patch logs metadata for the first unexpected remote read of each
scanrange when the flag fs_trace_remote_reads is set to true. This
logging is intended to help diagnose the root cause of remote reads.

Since a message may be logged for each scan range, there could be
several hundred lines of output in a degenerate case. However, the
remote read condition is not expected and verbose output may be needed
to diagnose the root cause.

Reviewed-by: Aman Sinha 
Change-Id: I8c6a3e92f44813048022edf2b91299b3b0a20257
Reviewed-on: http://gerrit.cloudera.org:8080/17062
Reviewed-by: Impala Public Jenkins 
Tested-by: Impala Public Jenkins 


> Add tracing for remote block reads
> --
>
> Key: IMPALA-10504
> URL: https://issues.apache.org/jira/browse/IMPALA-10504
> Project: IMPALA
>  Issue Type: Improvement
>Reporter: Kurt Deschler
>Assignee: Kurt Deschler
>Priority: Major
>
> While chasing performance issues, there were a large number of remote block 
> read messages in the logs. Need tracing to track down the source of these. 
> {noformat}
> Errors: Read 3.07 GB of data across network that was expected to be local.
> Block locality metadata for table 'tpcds_600_parquet.store_sales' may be 
> stale.
> This only affects query performance and not result correctness.
> One of the common causes for this warning is HDFS rebalancer moving some of 
> the file's blocks.
> If the issue persists, consider running "INVALIDATE METADATA 
> `tpcds_600_parquet`.`store_sales`"{noformat}
>  



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

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



[jira] [Commented] (IMPALA-10504) Add tracing for remote block reads

2021-02-12 Thread Kurt Deschler (Jira)


[ 
https://issues.apache.org/jira/browse/IMPALA-10504?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17283862#comment-17283862
 ] 

Kurt Deschler commented on IMPALA-10504:


[http://gerrit.cloudera.org:8080/17062]

> Add tracing for remote block reads
> --
>
> Key: IMPALA-10504
> URL: https://issues.apache.org/jira/browse/IMPALA-10504
> Project: IMPALA
>  Issue Type: Improvement
>Reporter: Kurt Deschler
>Assignee: Kurt Deschler
>Priority: Major
>
> While chasing performance issues, there were a large number of remote block 
> read messages in the logs. Need tracing to track down the source of these. 
> {noformat}
> Errors: Read 3.07 GB of data across network that was expected to be local.
> Block locality metadata for table 'tpcds_600_parquet.store_sales' may be 
> stale.
> This only affects query performance and not result correctness.
> One of the common causes for this warning is HDFS rebalancer moving some of 
> the file's blocks.
> If the issue persists, consider running "INVALIDATE METADATA 
> `tpcds_600_parquet`.`store_sales`"{noformat}
>  



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

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