yugabyte / tpcc

Repo to run TPCC benchmarks against YugabyteDB
Other
24 stars 21 forks source link

Use multi-column (warehouse_id, district_id) hash for tables #15

Closed ndeodhar closed 4 years ago

ndeodhar commented 4 years ago

This will better distribute the load. Today, since we only use warehouse_id as the hash, all data for a warehouse hits a single shard creating a hotspot. Note that all queries always provide warehouse_id and district_id. So, this is a safe change to make.