[
https://issues.apache.org/jira/browse/CASSANDRA-18709?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17757078#comment-17757078
]
Ekaterina Dimitrova commented on CASSANDRA-18709:
-------------------------------------------------
bq. I think something must be wrong with generate.sh and 3.x branches because
I don't get the option to run repeated tests
Repeated tests config should be added manually for Python DTests. Changed tests
are not detected automatically as Python DTests do not even reside in the C*
repo.
> Test failure: dtest.jmx_test.TestJMX.test_compactionstats
> ---------------------------------------------------------
>
> Key: CASSANDRA-18709
> URL: https://issues.apache.org/jira/browse/CASSANDRA-18709
> Project: Cassandra
> Issue Type: Bug
> Components: Test/dtest/python
> Reporter: Ekaterina Dimitrova
> Assignee: Brandon Williams
> Priority: Normal
> Fix For: 3.0.30, 3.11.16, 4.0.12, 4.1.4, 5.0
>
>
> Seen here:
> [https://ci-cassandra.apache.org/job/Cassandra-trunk/1646/testReport/dtest.jmx_test/TestJMX/test_compactionstats/]
> h3.
> {code:java}
> Error Message
> AttributeError: 'NoneType' object has no attribute 'named'
> Stacktrace
> self = <jmx_test.TestJMX object at 0x7fe78cfc2940> def
> test_compactionstats(self): """ @jira_ticket CASSANDRA-10504 @jira_ticket
> CASSANDRA-10427 Test that jmx MBean used by nodetool compactionstats properly
> updates the progress of a compaction """ cluster = self.cluster
> cluster.populate(1) node = cluster.nodelist()[0] cluster.start() # Run a
> quick stress command to create the keyspace and table node.stress(['write',
> 'n=1', 'no-warmup']) # Disable compaction on the table
> node.nodetool('disableautocompaction keyspace1 standard1')
> node.nodetool('setcompactionthroughput 1') node.stress(['write', 'n=150K',
> 'no-warmup']) node.flush() # Run a major compaction. This will be the
> compaction whose # progress we track. node.nodetool_process('compact') # We
> need to sleep here to give compaction time to start # Why not do something
> smarter? Because if the bug regresses, # we can't rely on jmx to tell us that
> compaction started. time.sleep(5) compaction_manager = make_mbean('db',
> type='CompactionManager') with JolokiaAgent(node) as jmx: progress_string =
> jmx.read_attribute(compaction_manager, 'CompactionSummary')[0] # Pause in
> between reads # to allow compaction to move forward time.sleep(2)
> updated_progress_string = jmx.read_attribute(compaction_manager,
> 'CompactionSummary')[0] var = 'Compaction@{uuid}(keyspace1, standard1,
> {progress}/{total})bytes' if self.cluster.version() >= LooseVersion('4.0'): #
> CASSANDRA-15954 var = 'Compaction({taskUuid}, {progress} / {total}
> bytes)@{uuid}(keyspace1, standard1)' > progress = int(parse.search(var,
> progress_string).named['progress']) E AttributeError: 'NoneType' object has
> no attribute 'named' jmx_test.py:218: AttributeError
> {code}
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
