[
https://issues.apache.org/jira/browse/CASSANDRA-6835?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13931206#comment-13931206
]
Benedict commented on CASSANDRA-6835:
-------------------------------------
Uploaded patch
[here|https://github.com/belliottsmith/cassandra/commits/iss-6835]
This actually makes a lot more changes than planned, a couple of which are
pretty important:
# It _fixes counter reads_ - they've been hitting the non-counter table since
this stress was introduced, which is kind of not the point
# Super columns reads had the same problem
# As part of this fix, I rescind the ability to specify a CF name, as it
doesn't really make much sense, and it only overrode the CF name for
non-counter, non-supercolumn operations. Making it more generic seemed like too
much work for the payoff
# It permits operations on counters to operate over a variable number of
columns, selecting a random sample of the possible column names (note that
reads may still fail if they get nothing back, so ideally all possible columns
should be populated once before any random read/write workload is let loose)
# It permits varying the amount a counter is incremented by, based on a
distribution
# It permits selecting if you want to perform a range slice query (/select *)
or a name filter query for reads (defaulting to the latter where possible)
# It slightly modifies the -mode parameter spec to make it clearer what kind of
CQL3/2 connection you're making
> cassandra-stress should support a variable number of counter columns
> --------------------------------------------------------------------
>
> Key: CASSANDRA-6835
> URL: https://issues.apache.org/jira/browse/CASSANDRA-6835
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Benedict
> Assignee: Benedict
> Priority: Minor
>
--
This message was sent by Atlassian JIRA
(v6.2#6252)
