Yida Wu has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/23942 )

Change subject: IMPALA-14230: Add catch-up mode for event processing
......................................................................


Patch Set 5:

(5 comments)

http://gerrit.cloudera.org:8080/#/c/23942/4//COMMIT_MSG
Commit Message:

http://gerrit.cloudera.org:8080/#/c/23942/4//COMMIT_MSG@20
PS4, Line 20: trigger heavy
            : operations,
> This always mean loading/refreshing tables/partitions, or there are also ot
Yes, I think these are operations we want to avoid, and maybe the local 
operations that involve lock stuff. Added commit message here.


http://gerrit.cloudera.org:8080/#/c/23942/4/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/23942/4/fe/src/main/java/org/apache/impala/catalog/events/MetastoreEvents.java@1242
PS4, Line 1242:       long lag = evaluateCatchUpLag();
              :       if (lag > 0) {
              :         if (catalog_.i
> I don't understand this "if".
Yeah, I was hesitated for this because maybe it is not that wrong to log when 
the table is already invalidated even though it is not invalidated by this 
action.
I've updated invalidateTableIfExists() to return null when skipIfIncomplete is 
true and the table is already incomplete. This will make the logging more 
correct.


http://gerrit.cloudera.org:8080/#/c/23942/4/fe/src/main/java/org/apache/impala/catalog/events/MetastoreEvents.java@1978
PS4, Line 1978:
> Is it a good idea to evaluate this before isSelfEvent() and isOlderEvent()?
Added some comments.


http://gerrit.cloudera.org:8080/#/c/23942/4/tests/custom_cluster/test_events_custom_configs.py
File tests/custom_cluster/test_events_custom_configs.py:

http://gerrit.cloudera.org:8080/#/c/23942/4/tests/custom_cluster/test_events_custom_configs.py@728
PS4, Line 728: onfigured thresh
> Is there a reason for creating the table in Hive?
Thought transactional table needs hive, but it seems not necessary. Changed it 
to impala.


http://gerrit.cloudera.org:8080/#/c/23942/4/tests/custom_cluster/test_events_custom_configs.py@760
PS4, Line 760:                   
"--debug_actions=catalogd_event_processing_delay:SLEEP@3
> Shouldn't the test also check if the given event type invalidates the table
That makes sense. Made the change to catch invalidated logs.



--
To view, visit http://gerrit.cloudera.org:8080/23942
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: Ib906c06346d5d3159999eeac632e1318bc060065
Gerrit-Change-Number: 23942
Gerrit-PatchSet: 5
Gerrit-Owner: Yida Wu <[email protected]>
Gerrit-Reviewer: Anonymous Coward <[email protected]>
Gerrit-Reviewer: Csaba Ringhofer <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Quanlong Huang <[email protected]>
Gerrit-Reviewer: Sai Hemanth Gantasala <[email protected]>
Gerrit-Reviewer: Yida Wu <[email protected]>
Gerrit-Comment-Date: Tue, 24 Feb 2026 21:02:12 +0000
Gerrit-HasComments: Yes

Reply via email to