Quanlong Huang has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/19155 )

Change subject: IMPALA-11626: Handle COMMIT_COMPACTION_EVENT from HMS
......................................................................


Patch Set 9:

(4 comments)

http://gerrit.cloudera.org:8080/#/c/19155/9/fe/src/main/java/org/apache/impala/catalog/events/MetastoreEvents.java
File fe/src/main/java/org/apache/impala/catalog/events/MetastoreEvents.java:

http://gerrit.cloudera.org:8080/#/c/19155/9/fe/src/main/java/org/apache/impala/catalog/events/MetastoreEvents.java@2681
PS9, Line 2681:            "Commit compaction event", 
FileMetadataLoadOpts.FORCE_LOAD);
If the event is triggered by a table level compaction, partitionName_ will be 
DEFAULT_PARTITION_NAME. I think for partitioned tables, partition of this name 
doesn't exist. Only non-partitioned table has such a partition. To trigger full 
table refresh for table-level events, I think we need reloadTableFromCatalog() 
or something similar.


http://gerrit.cloudera.org:8080/#/c/19155/9/fe/src/main/java/org/apache/impala/service/CatalogOpExecutor.java
File fe/src/main/java/org/apache/impala/service/CatalogOpExecutor.java:

http://gerrit.cloudera.org:8080/#/c/19155/9/fe/src/main/java/org/apache/impala/service/CatalogOpExecutor.java@4456
PS9, Line 4456:     if (table == null && !(table instanceof IncompleteTable) &&
              :         (table instanceof HdfsTable)) {
hmm, if 'table' is null, it can't be an instance of HdfsTable. This condition 
is always false..

Do you want to use "table != null"? If so, "table instanceof HdfsTable" already 
implies 'table' is not null and not an instance of IncompleteTable.


http://gerrit.cloudera.org:8080/#/c/19155/9/fe/src/test/resources/hive-site.xml.py
File fe/src/test/resources/hive-site.xml.py:

http://gerrit.cloudera.org:8080/#/c/19155/9/fe/src/test/resources/hive-site.xml.py@167
PS9, Line 167:    'hive.compactor.delta.num.threshold': 1,
Could you add a comment about why we need this for all the tests?

This will also be used in data loading. Will it impact the data loading 
performance?


http://gerrit.cloudera.org:8080/#/c/19155/9/tests/custom_cluster/test_events_custom_configs.py
File tests/custom_cluster/test_events_custom_configs.py:

http://gerrit.cloudera.org:8080/#/c/19155/9/tests/custom_cluster/test_events_custom_configs.py@327
PS9, Line 327:     assert EventProcessorUtils.get_event_processor_status() == 
"ACTIVE"
We also need tests on unpartitioned tables and table level compaction on 
partitioned tables.



--
To view, visit http://gerrit.cloudera.org:8080/19155
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I464faedb4e3bbcd417bab2e3cb0d57e339d42605
Gerrit-Change-Number: 19155
Gerrit-PatchSet: 9
Gerrit-Owner: Sai Hemanth Gantasala <[email protected]>
Gerrit-Reviewer: Daniel Becker <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Quanlong Huang <[email protected]>
Gerrit-Reviewer: Sai Hemanth Gantasala <[email protected]>
Gerrit-Comment-Date: Wed, 15 Feb 2023 01:12:45 +0000
Gerrit-HasComments: Yes

Reply via email to