Tim Armstrong has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/13933 )

Change subject: IMPALA-8627: Enable catalog-v2 in tests
......................................................................


Patch Set 11:

(4 comments)

I think this generally makes sense. I had a couple of questions about the test 
and a suggestion to avoid some code duplication.

http://gerrit.cloudera.org:8080/#/c/13933/11/tests/hs2/hs2_test_suite.py
File tests/hs2/hs2_test_suite.py:

http://gerrit.cloudera.org:8080/#/c/13933/11/tests/hs2/hs2_test_suite.py@74
PS11, Line 74:     def add_session(self, cluster_properties):
I think we can avoid all of the duplication with the above add session function 
if we define a helper function that receives all the required arguments and the 
function to invoke in the middle, i.e.

  def add_session_helper(self, protocol_version, conf_overlay, close_session, 
fn):
    ...
    fn()
    ...

Then we could invoke it with something like

   add_session_helper(self, protocol_version, conf_overlay, close_session, 
lambda: fn(self, cluster_properties))

I tried it out here - seems to work: 
https://github.com/timarmstrong/impala/tree/refactor-session


http://gerrit.cloudera.org:8080/#/c/13933/11/tests/hs2/test_hs2.py
File tests/hs2/test_hs2.py:

http://gerrit.cloudera.org:8080/#/c/13933/11/tests/hs2/test_hs2.py@440
PS11, Line 440:   @needs_session_cluster_properties()
If we're defining a new decorator anyway, could we pass in unique_database too 
and remove execute_serially? That was a workaround for the decorator issue, but 
it looks like you might have figured it out.


http://gerrit.cloudera.org:8080/#/c/13933/11/tests/query_test/test_observability.py
File tests/query_test/test_observability.py:

http://gerrit.cloudera.org:8080/#/c/13933/11/tests/query_test/test_observability.py@318
PS11, Line 318:         r'CatalogFetch.Config.Misses|CatalogFetch.Config.Hits',
Is it guaranteed that we always get one or the other?


http://gerrit.cloudera.org:8080/#/c/13933/11/tests/query_test/test_observability.py@340
PS11, Line 340:     self.__verify_profile_event_sequence(load_event_regexes, 
runtime_profile)
Do we need to test that the above regexes are in sequential lines, or just that 
they're present? Originally this was meant to assert that the events were in 
the specific order.

I guess the advantage of this is that we'll know to update the test when new 
metrics are added. So seems OK.



--
To view, visit http://gerrit.cloudera.org:8080/13933
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: Iddbde666de2b780c0e40df716a9dfe54524e092d
Gerrit-Change-Number: 13933
Gerrit-PatchSet: 11
Gerrit-Owner: Vihang Karajgaonkar <[email protected]>
Gerrit-Reviewer: Bharath Vissapragada <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Tim Armstrong <[email protected]>
Gerrit-Reviewer: Vihang Karajgaonkar <[email protected]>
Gerrit-Comment-Date: Tue, 06 Aug 2019 06:30:39 +0000
Gerrit-HasComments: Yes

Reply via email to