cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Stefania (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-10753) Fix completion problems breaking clqshlib tests
Date Fri, 27 Nov 2015 03:02:11 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-10753?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15029366#comment-15029366
] 

Stefania commented on CASSANDRA-10753:
--------------------------------------

Thanks for your review:

bq.  In test_timestamp_output since the TZ property is ignored I guess we can just remove
the second equals statement which is the same as the first. What do you think?

Agreed, removed.    

bq. The max_window_size_seconds problem is also affecting 2.2, could you provide a 2.2 patch?
(It's not on jenkins but since we'll still maintain 2.2 for some time, I guess we should fix
it).

To fix 2.2 I think we need the entire patch since according to [~aholmber] [comment|https://issues.apache.org/jira/browse/CASSANDRA-9302?focusedCommentId=15027191&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15027191]
on CASSANDRA-9302 we should be running the current driver. So I backported the entire patch
to 2.2 (via cherry-pick of individual commits) and then upgraded the bundled driver to the
same as 3.0. By the way, do you know which version it is? I could not even find commit {{3f15725}}
on any branch in the driver repository. Shouldn't we upgrade to a later and more official
version instead?

I also started noticing some timeout failures in the cqlshlib tests, on all branches, when
dropping keyspaces. I did not notice them before and I think it's because I just migrated
to a loan laptop where I installed Linux on an hybrid HDD rather than SSD. It takes a crazy
long time to flush all tables before the final KEYSPACE DROPPED notification is sent to the
driver. For 2.2 I needed at least 20 seconds but for 3.0+ it took up to 35 seconds so I increased
the timeout to 45 seconds in 3.0+. Do you think this is normal?

Here are the patches and CI, but CASSCI is currently down so I cannot start any jobs:

||2.2||3.0||trunk||
|[patch|https://github.com/stef1927/cassandra/commits/10753-2.2]|[patch|https://github.com/stef1927/cassandra/commits/10753-3.0]|[patch|https://github.com/stef1927/cassandra/commits/10753]|
|[dtest|http://cassci.datastax.com/view/Dev/view/stef1927/job/stef1927-10753-2.2-dtest/]|[dtest|http://cassci.datastax.com/view/Dev/view/stef1927/job/stef1927-10753-3.0-dtest/]|[dtest|http://cassci.datastax.com/view/Dev/view/stef1927/job/stef1927-10753-dtest/]|

I verified the cqlshlib tests locally on all branches. I'll start the dtests as soon as CASSCI
is back.


> Fix completion problems breaking clqshlib tests
> -----------------------------------------------
>
>                 Key: CASSANDRA-10753
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-10753
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Stefania
>            Assignee: Stefania
>             Fix For: 3.0.x
>
>
> The recent changes in the python driver APIs have caused some completion problems as
indicated by 2 of these failing tests:
> http://cassci.datastax.com/job/trunk_cqlshlib/579/testReport/
> The third failing test, {{test_timestamp_output}}, has been failing for some time due
to uncertainty on what to do regarding timezone conversion but it too can be changed to reflect
the fact that we convert the timestamp to UTC.
> Finally, {{max_window_size_seconds}} was recently added to the compaction properties
and this caused 2 more tests to fail. It cannot be seen on Jenkins because of the relative
import problem introduced by CASSANDRA-9304. 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message