Closed sushanb closed 2 years ago
I don't think we supports preprovisioned databases at the moment in general. Brandon may have more context.
Hi sorry for the delayed response. We don't have existing functionality to do this, however it could be implemented with some effort. I recommend taking a look at linux_packages/hbase.py as well as linux_benchmarks/cloud_bigtable_ycsb_benchmark.py, which is able to use preprovisioned bigtable instances.
Closing this now, feel free to reopen.
For the context, ycsb allows to do this through hbase-site.xml file in the class path.
bin/ycsb load hbase10 -P workloads/workloada -cp /home/ycsb-hbase/conf -p table=usertable -p columnfamily=family -threads 100.
Is there a way to do this in perfkitbenchmarker? As i see, the script does the hbase installation and everything. I wonder how hard it will be to force it to use the hbase-site and run the workload?