hs-benchmark - a simple HammerServer benchmark
hs-benchmark URL DURATION
The benchmark is run for the stated duration (seconds). Data are inserted and retrieved, the HammerServer is expected at the stated URL. The output consists of lines, that state:
The inserted string (in the format Test-data-00000001, Test-data-00000002 and so on);
Duration of the insertion
Duration of the retrieval
Total duration
Finally the script reports the avarages of the above timings.
The purpose of the benchmark is that simple inserts and retrievals can be timed very easily. Also, the benchmark can be left to run for a long time (e.g., overnight, or for a week) to see whether the timings degrade with database usage.
Insertions and retrievals are single transactions, i.e., separate calls to the HammerServer. Upon retrieval, hs-benchmark checks that the retrieved text is the same as the inserted text.
Note that the stated times include all traffic overhead, and are therefore not ``pure'' server-side timings. If you want to see the server-side durations, enable the PERF logging in the configuration file etc/hammerserver.cfg, and monitor the log log/hammerserverd.log.
For a discussion of the configuration (notably performance logging), see hsctl.
The tamper-resistant server, this benchmark script, all tooling, and the documentation were written by Karel Kubat / karel@e-tunity.com. Copyright (c) 2009 ff. Distributed under GPLV3.