[Impala-ASF-CR] IMPALA-9903: Reduce Kudu openTable calls per query

["Tim Armstrong (Code Review)" <ge...@cloudera.org>]

Tim Armstrong has posted comments on this change. ( http://gerrit.cloudera.org:8080/16120 )

Change subject: IMPALA-9903: Reduce Kudu openTable calls per query
......................................................................


Patch Set 2:

(3 comments)

http://gerrit.cloudera.org:8080/#/c/16120/2/fe/src/main/java/org/apache/impala/catalog/FeKuduTable.java
File fe/src/main/java/org/apache/impala/catalog/FeKuduTable.java:

http://gerrit.cloudera.org:8080/#/c/16120/2/fe/src/main/java/org/apache/impala/catalog/FeKuduTable.java@166
PS2, Line 166:     public static TResultSet getPartitions(FeKuduTable table)
These are methods that implement the show partitions DDL, so we don't need to worry about caching these - I think to maintain the current semantics we need to open the table once per DDL anyway.

I.e. I don't think we want to change the behaviour of these methods. Maybe we can be a comment to be explicit that we want to fetch the latest version of the Kudu table.


http://gerrit.cloudera.org:8080/#/c/16120/2/fe/src/main/java/org/apache/impala/catalog/local/LocalKuduTable.java
File fe/src/main/java/org/apache/impala/catalog/local/LocalKuduTable.java:

http://gerrit.cloudera.org:8080/#/c/16120/2/fe/src/main/java/org/apache/impala/catalog/local/LocalKuduTable.java@56
PS2, Line 56:   private org.apache.kudu.client.KuduTable kuduTable_;
Caching it in LocalTable makes sense since it's per-query anyway. So this part is fine - caching it here means we can open it once per query then re-use it.


http://gerrit.cloudera.org:8080/#/c/16120/2/fe/src/main/java/org/apache/impala/planner/KuduScanNode.java
File fe/src/main/java/org/apache/impala/planner/KuduScanNode.java:

http://gerrit.cloudera.org:8080/#/c/16120/2/fe/src/main/java/org/apache/impala/planner/KuduScanNode.java@135
PS2, Line 135:       org.apache.kudu.client.KuduTable rpcTable = kuduTable_.getKuduTable();
I think this invocation should go via 'analyzer' to retrieve the per-query cached version of the table. That would work for both catalog implementations.



-- 
To view, visit http://gerrit.cloudera.org:8080/16120
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: Iec12a5be9b30e19a123142af5453a91bd4300b63
Gerrit-Change-Number: 16120
Gerrit-PatchSet: 2
Gerrit-Owner: Grant Henke <gr...@apache.org>
Gerrit-Reviewer: Grant Henke <gr...@apache.org>
Gerrit-Reviewer: Impala Public Jenkins <im...@cloudera.com>
Gerrit-Reviewer: Tim Armstrong <ta...@cloudera.com>
Gerrit-Reviewer: Vihang Karajgaonkar <vi...@cloudera.com>
Gerrit-Comment-Date: Fri, 17 Jul 2020 17:33:51 +0000
Gerrit-HasComments: Yes