I can confirm that those noisy novnode_dtest failures are from the Python
driver problem I tracked down. I forgot to trigger a new CI run after that
was fixed. Today is down to one failure, which is much closer to what we
expect.

On Wed, Aug 3, 2016 at 7:45 PM, Josh McKenzie <jmcken...@apache.org> wrote:

> Summary:
>    Somewhat messy day, but I believe things should look less noisy tomorrow
> so
>    we can get a better view of what we have left to a stable board. Going
> to try
>    a slightly new format:
>
> [Failures]
>    testall:       1
>    dtest:         0
>    novnode_dtest: 17
>    upgrade:       110
>
> [Details]
>    testall (1):
>       CommitLogSegmentManagerTest.testCompressedCommitLogBackpressure
>          CASSANDRA-12283, currently assigned to blerer
>
>    novnode_dtest (17):
>       Bulk: Hard to separate signal from noise due to rampant timeouts and
> nodes
>          being marked down. This is possibly due to a problem in the python
> driver
>          that was tracked down today - Joel, can you confirm?
>          An example message:
>             ('Unable to complete the operation against any hosts', {<Host:
> 127.0.0.3
>             datacenter1>: ConnectionException('Host has been marked down or
> removed',),
>             <Host: 127.0.0.2 datacenter1>: ConnectionException('Host has
> been marked down
>             or removed',), <Host: 127.0.0.1 datacenter1>:
> ConnectionException('Host has
>             been marked down or removed',)})
>       cql_tests.MiscellaneousCQLTester.prepared_statement_
> invalidation_test
>          **Needs Triage**
>          New failure as of yesterday, tracked by CASSANDRA-12361
>       materialized_views_test.TestMaterializedViews.complex_repair_test
>          Assigned to cyeksigian
>          Intermittent failure
>          Last comment from Philip indicates we want to move this to
> dtest_large. Pinged Jim.
>
>    upgrade (110):
>       CASSANDRA-12236 was committed today so that should help with the
> noise
>       Sylvain commented on CASSANDRA-10848 yesterday concerning some
> potential
>          problems with paging tests and whether or not we're setting the
> protocol
>          version for those tests. Should be interesting to see how things
> look
>          after the run with 12236 in.
>
> ~Josh
>



-- 

<http://www.datastax.com/>

Joel Knighton
Cassandra Developer | joel.knigh...@datastax.com

<https://www.linkedin.com/company/datastax>
<https://www.facebook.com/datastax> <https://twitter.com/datastax>
<https://plus.google.com/+Datastax/about>
<http://feeds.feedburner.com/datastax> <https://github.com/datastax/>

<http://cassandrasummit.org/Email_Signature>

Reply via email to