[
https://issues.apache.org/jira/browse/CASSANDRA-6572?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14154440#comment-14154440
]
Jonathan Shook commented on CASSANDRA-6572:
-------------------------------------------
I believe that we do need a tool to help characterize real workloads. There is
a distinct difference between a suggestive combination tests and something that
is based on the access patterns, timings, data distribution, etc of a real
working system. While the new stress tool is nifty, it is a stress tool, not a
workload characterization tool. There are many systems in use which have access
patterns that are not trivially easy to test with a service mix. Also,
understanding the behavior of clients and the cluster in-situ can be very
helpful in various ways-- troubleshooting, capacity planning, scale testing,
etc. Lots of time is spent just getting users to the point where they
understand what their system is doing. A proper sample of real behavior is the
best possible data to start from.
I'm in favor of keeping some form of real workload sampling on the table. Even
if the stress tool is enhanced, there is still a workload characterization
problem to solve here, which I believe is true to the spirit of the original
request.
> Workload recording / playback
> -----------------------------
>
> Key: CASSANDRA-6572
> URL: https://issues.apache.org/jira/browse/CASSANDRA-6572
> Project: Cassandra
> Issue Type: New Feature
> Components: Core, Tools
> Reporter: Jonathan Ellis
> Assignee: Carl Yeksigian
> Fix For: 2.1.1
>
> Attachments: 6572-trunk.diff
>
>
> "Write sample mode" gets us part way to testing new versions against a real
> world workload, but we need an easy way to test the query side as well.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
