Vuk Ercegovac has posted comments on this change. ( http://gerrit.cloudera.org:8080/9538 )
Change subject: IMPALA-6602: fixes flaky expiration test ...................................................................... Patch Set 1: (3 comments) http://gerrit.cloudera.org:8080/#/c/9538/1/tests/custom_cluster/test_query_expiration.py File tests/custom_cluster/test_query_expiration.py: http://gerrit.cloudera.org:8080/#/c/9538/1/tests/custom_cluster/test_query_expiration.py@80 PS1, Line 80: assert (client.get_state(short_timeout_expire_handle) == > These may have the same problem too, right? my thought here is that we think we wait enough so that we're just looking up the current state (e.g., we're not waiting until a condition switches). so that's why I didn't guard these. http://gerrit.cloudera.org:8080/#/c/9538/1/tests/custom_cluster/test_query_expiration.py@159 PS1, Line 159: assert client.get_state(timeout_handle) == client.QUERY_STATES['EXCEPTION'] > These asserts have the same problem, right? same thought process as in the previous comment. primarily, I looked for the wait_for_metric_value. http://gerrit.cloudera.org:8080/#/c/9538/1/tests/custom_cluster/test_query_expiration.py@175 PS1, Line 175: def __expect_client_state(self, client, handle, expected_state, max_tries=10): > Should we have a time bound instead of number of tries? If get_state() was we're doing an rpc for this one so I thought it should be slow enough. since there are timing assumptions around the call sites, I thought it may make this more difficult to reason with. let me know if you're more comfortable with a time bound. -- To view, visit http://gerrit.cloudera.org:8080/9538 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I7aabed87d84d5cfd8078cc6c39df48e22ff30afc Gerrit-Change-Number: 9538 Gerrit-PatchSet: 1 Gerrit-Owner: Vuk Ercegovac <[email protected]> Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-Reviewer: Vuk Ercegovac <[email protected]> Gerrit-Comment-Date: Wed, 07 Mar 2018 22:41:46 +0000 Gerrit-HasComments: Yes
