search

GoogleCloudPlatform / PerfKitBenchmarker

PerfKit Benchmarker (PKB) contains a set of benchmarks to measure and compare cloud offerings. The benchmarks use default settings to reflect what most users will see. PerfKit Benchmarker is licensed under the Apache 2 license terms. Please make sure to read, understand and agree to the terms of the LICENSE and CONTRIBUTING files before proceeding.
https://googlecloudplatform.github.io/PerfKitBenchmarker/
Apache License 2.0
1.91k stars 512 forks source link

YCSB db size too small. #552

Open yuyantingzero opened 9 years ago

yuyantingzero commented 9 years ago

Default ycsb config for workloadb (95% read, 5% update) on n1-standard-8 doesn't issue disk reads with newer cassandra version (2.1.10), the entire db fit in memory.

kivio commented 9 years ago

@yuyantingzero I fix this problem by change workload file, you can change there number of columns in each record, size of column and number of rows (db size).

cmccoy commented 9 years ago

The current configuration populates the database with ~1GB of records. This is certainly too small for stressing Cassandra - I usually increase this to at least 2x the total memory of the cluster (via --ycsb_recordcount).

I'm OK increasing the default, but for in-memory cases (Aerospike if it's not configured to use disk, Redis) we need to keep the database size relatively small. How about for MongoDB, Aerospike (disk only), HBase, and Cassandra increasing the default recordcount to 2x the total memory of the data serving nodes?