Joe McDonnell has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/22231 )

Change subject: IMPALA-13620: Refresh compute_table_stats.py script
......................................................................


Patch Set 9:

Ok, I did a little bit of digging, and I think our theory isn't quite right. 
There is a postgres connection limit set to 100, but we aren't hitting it. HMS 
has a Hikari connection pool that it uses to connect to postgres and that has a 
size limit. By default, the connection pool has a size limit of 10. There are 
two connection pools for different things, so that's a max of 20 connections to 
postgres. The Hikari connection pool has a connection timeout, which is how 
long it will wait to get a connection from the pool. It defaults to 30 seconds, 
and that is what we are hitting. The good news is that these Hikari settings 
are configurable by modifying hive-site.xml.

We can change the size of the connection pool by setting 
datanucleus.connectionPool.maxPoolSize. I think we could try bumping this to 
20-30. See 
https://github.com/apache/hive/blob/master/standalone-metastore/metastore-common/src/main/java/org/apache/hadoop/hive/metastore/conf/MetastoreConf.java#L695-L702

The 30 second connectionTimeout is also configurable by setting 
hikaricp.connectionTimeout (units are milliseconds) (based on 
https://github.com/apache/hive/blob/master/standalone-metastore/metastore-server/src/main/java/org/apache/hadoop/hive/metastore/datasource/HikariCPDataSourceProvider.java#L43
 and 
https://github.com/brettwooldridge/HikariCP?tab=readme-ov-file#frequently-used 
). I don't think there are major issues from doubling or tripling that.


--
To view, visit http://gerrit.cloudera.org:8080/22231
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I1ebf02f95b957e7dda3a30622b87e8fca3197699
Gerrit-Change-Number: 22231
Gerrit-PatchSet: 9
Gerrit-Owner: Riza Suminto <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Jason Fehr <[email protected]>
Gerrit-Reviewer: Joe McDonnell <[email protected]>
Gerrit-Reviewer: Quanlong Huang <[email protected]>
Gerrit-Reviewer: Riza Suminto <[email protected]>
Gerrit-Comment-Date: Mon, 06 Jan 2025 23:35:41 +0000
Gerrit-HasComments: No

Reply via email to