CH3 measures a mixed workload performance with transactional, analytical and full test search (FTS) queries in Couchbase Server. However, it is not specific to Couchbase, rather it can be implemented on other NoSQL platforms that support query, analytics and search services.
** For Couchbase, we have implemented "nestcollectionsdriver.py". Create a new file in the 'drivers' directory for your system that follows the proper naming convention. For example, the driver class in "nestcollectionsdriver.py" is 'NestcollectionsDriver'.
*** CH3 supports FTS queries, transactional queries based on TPC-C, and TPC-H-equivalent analytical queries.
In the following commands, replace the IP addresses with the appropriate ones. Besides, the parameters we use are:
userid = Administrator
password = password
warehouses = 10
aclients = 1
fclients = 40
tclients = 40
- Create the "bench" bucket with scope and collections
sh -x util/cbcrbucketcollection.sh <Data_IPAddress> Administrator:password
- Create GSI indexes
/opt/couchbase/bin/cbq -e=<Index_IPAddress>:8093 -u=Administrator -p=password < util/cbcrindexcollection.sql
- Prepare shadow for analytics
/opt/couchbase/bin/cbq -e=<Analytics_IPAddress>:8095 -u=Administrator -p=password < util/crdataset.n1ql
/opt/couchbase/bin/cbq -e=<Analytics_IPAddress>:8095 -u=Administrator -p=password < util/crdatasetindexes.n1ql
- Create FTS indexes
sh -x util/cbcrftsindexcollection.sh <FTS_IPAddress> Administrator:password
- Load data into "bench" bucket
python3 ./tpcc.py --warehouses 10 --no-execute nestcollections \
--data-url <Data_IPAddress> --userid Administrator --password password \
--datasvc-bulkload --tclients 40
Note that we have three load modes: (1) datasvc-bulkload, (2) datasvc-load, (3) qrysvc-load. In the above command, we have "nestcollections" for the driver name.
- Execute CH3 benchmark with appropriate number of Txn, FTS and analytical clients
python3 ./tpcc.py --warehouses 10 --aclients 1 --tclients 40 --fclients 40 \
--query-url <Query_IPAddress>:8093 --analytics-url <Analytics_IPAddress>:8095 --fts-url <FTS_IPAddress>:8094 \
--query-iterations 2 --warmup-query-iterations 1 nestcollections \
--userid Administrator --password password --no-load
If there is no analytical clients, then we need "duration" parameter.
python3 ./tpcc.py --warehouses 10 --tclients 40 --aclients 0 --fclients 40 \
--query-url <Query_IPAddress>:8093 --analytics-url <Analytics_IPAddress>:8095 --fts-url <FTS_IPAddress>:8094 \
--duration 180 --warmup-duration 90 nestcollections \
--userid Administrator --password password --no-load
CH3 extends CH2 benchmark framework.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent