[
https://issues.apache.org/jira/browse/CASSANDRA-16342?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Yifan Cai updated CASSANDRA-16342:
----------------------------------
Test and Documentation Plan:
Perf. Test Plan
# Run data prepopulation for both 3.0 and 4.0 clusters. 200GB per node.
# Run steady state load for X seconds and compare the compaction performance.
The workload used in the test was generated from tlp-stress.
Steady state load, Write : Delete == 4 : 1, and the QPS was kept at 1500/s.
Status: Patch Available (was: Open)
Result report link:
[https://github.com/yifan-c/CASSANDRA-15581-COMPACTION-TEST/blob/main/CASSANDRA-16342-16343/Compaction%20Perf%20Comparison%20Between%203.0%20and%204.0%20(STCS%20and%20LCS).pdf]
The compaction performance comparison mainly focuses on the compaction related
metrics, e.g. compaction throughput, pending tasks, the number of unleveled
sstables (LCS only), etc. Query latency is *not* compared, since is can be
affected by many other components that changed between 3.0 and 4.0.
Under the same load, the LCS compaction in 4.0 has better performance. The
compaction throughput is higher and the 4.0 cluster keeps up with the load.
However, the compaction in 3.0 cluster is lagging behind as the number of the
unleveled sstables increases during the test.
Details can be found in the report linked above.
> LCS steady state load of 4.0 vs. 3.x performance test
> -----------------------------------------------------
>
> Key: CASSANDRA-16342
> URL: https://issues.apache.org/jira/browse/CASSANDRA-16342
> Project: Cassandra
> Issue Type: Sub-task
> Components: Test/benchmark
> Reporter: Yifan Cai
> Assignee: Yifan Cai
> Priority: Normal
>
> The testing cluster should be pre-populated with ~200GB data in each node.
> Run the steady state workload to compare the read, write and compaction
> performance between baseline (3.x cluster) and 4.0 cluster.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
