TPC-H test report of OceanBase Database

This topic introduces the test environment, testing plan, and detailed test results for the TPC-H test.

Test environment

The deployment environment uses Alibaba Cloud Elastic Compute Service (ECS) as an example.

• Hardware configuration

  Service type	ECS type	Instance count	Number of cores	Memory
  OceanBase Database	ecs.r8i.8xlarge	3	32C	256G Each instance has a log disk of 400 GB and a data disk of 1200 GB, both with the performance level of PL1.

• Software versions

  Service type	Software version
  OceanBase Database	Enterprise Edition: OceanBase 4.3.5.0 Community Edition: OceanBase_CE 4.3.5.0
  TPC-H	V3.0.0
  OS	CentOS Linux release 7.9.2009 (Core)

• Tenant specifications

  obd cluster tenant create obperf -n tpcc_tenant --max-cpu=28 --memory-size=180G -–zone-list=zone1,zone2,zone3 -–primary-zone=RANDOM  --locality=F@zone1,F@zone2,F@zone3 --charset=utf8 -s 'ob_tcp_invited_nodes="%"' --optimize=htap
  

  Note

  In this example, --optimize=htap is the default workload type. In a production environment, select an appropriate workload type based on your cluster type.

Test solution

• Deploy an OceanBase Database cluster using obd. The TPC-H client must run on a single server as the stress machine. You do not need to deploy ODP. Directly connect to any server.

• The OceanBase cluster consists of three nodes with a 1:1:1 configuration. After the cluster is deployed, create a tenant and a user (sys is an built-in system tenant for cluster management. Do not use the sys tenant for testing.) for running the TPC-H test. Set the primary_zone of the tenant to RANDOM.

• The amount of test data is 100 GB.

• For more information, see Perform the TPC-H benchmark with OceanBase Database.

Test results

The TPC-H test results are for data sizes of 100 GB and 1 TB, covering hot run, first run, and cold run scenarios. The tests were conducted with 96 concurrent threads.