lijian-ruckus opened a new issue, #15635:
URL: https://github.com/apache/druid/issues/15635
Druid broker has high cpu usage for parsing certain sqls and fails to return.
### Affected Version
v2.6.1
### Description
Please include as much detailed information about the problem as possible.
- Cluster size: can be produced in a minimum setup with wikipedia data
This issue could be reproduced by the following query against the example
wikipedia data
`SELECT page, max("commentLength") as "max_comment" FROM "wikipedia" where
("__time" >= '2016-06-27 00:00:00' AND "__time" < '2016-06-27 09:00:00') AND
page='Kraft' AND page='1961' GROUP BY "page" HAVING ((MAX(commentLength) > 0))
ORDER BY "max_comment" DESC LIMIT 10
`
For the above query empty data is expected to return instead of a run-away
process hogging the cpu for ever.
In the impacted broker, the "top -H -p" command will show the process
hogging the cpu with a command sql[xxxxx and the query will never return.
To reproduce the issue the following conditions need to meet at the same
time:
1. A "HAVING" clause
3. more than 1 AND conditions on the same column (e.g page in the above
query)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
