Tim Armstrong has posted comments on this change. ( http://gerrit.cloudera.org:8080/9921 )
Change subject: IMPALA-6793: Fix empty metadata after statestore restarts ...................................................................... Patch Set 2: (2 comments) I don't understand the catalog part of this but took a look at the test. http://gerrit.cloudera.org:8080/#/c/9921/1/tests/custom_cluster/test_restart_services.py File tests/custom_cluster/test_restart_services.py: http://gerrit.cloudera.org:8080/#/c/9921/1/tests/custom_cluster/test_restart_services.py@41 PS1, Line 41: sleep(60) > This test looks flaky but I haven't found find a better way to do it. The b Is there a metric we can poll that would indicate the expected state? http://gerrit.cloudera.org:8080/#/c/9921/2/tests/custom_cluster/test_restart_services.py File tests/custom_cluster/test_restart_services.py: http://gerrit.cloudera.org:8080/#/c/9921/2/tests/custom_cluster/test_restart_services.py@38 PS2, Line 38: os.system("kill -9 $(pidof statestored)") We should be able to use the Process.restart() method (https://github.com/apache/impala/blob/master/tests/common/impala_cluster.py#L175), something like below. Just running the shell commands directly leaves the ImpalaCluster object in a weird state. self.cluster.statestored.restart() -- To view, visit http://gerrit.cloudera.org:8080/9921 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I332a60e172af84b93b3544373fe363cdced5e8d0 Gerrit-Change-Number: 9921 Gerrit-PatchSet: 2 Gerrit-Owner: Tianyi Wang <[email protected]> Gerrit-Reviewer: Dimitris Tsirogiannis <[email protected]> Gerrit-Reviewer: Tianyi Wang <[email protected]> Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-Comment-Date: Tue, 03 Apr 2018 23:44:31 +0000 Gerrit-HasComments: Yes
