I’m trying to build a simple performance test on a rails 3 app that
operates on a large database (~150GB). We copy production data to our
dev/test setup every night so that we have a realistic environment for
development.
I followed the guide on creating a test and tried to run it using
$ rake test:benchmark
Unbelievably, this tried to drop my test database!!! Fortunately it
didn’t
have permission to do this and failed. Is there a way to skip this step
and just run the test against the DB as it exists already. From the
guide,
it seems like most rails apps have their test data in yaml files or
something but I don’t understand how this would scale to a production
environment. Code that works well on 100 or so database rows is not
necessarily going to work well on millions of rows.
Feel free to point me in the direction of relevant documentation and
tell
me to RTFM.
Unbelievably, this tried to drop my test database!!! Fortunately it didn’t
have permission to do this and failed.
Is there a way to skip this step and
just run the test against the DB as it exists already. From the guide, it
seems like most rails apps have their test data in yaml files or something
but I don’t understand how this would scale to a production environment.
Code that works well on 100 or so database rows is not necessarily going to
work well on millions of rows.
The test environment is designed for testing that stuff works, not for
performance testing. In order for tests to be repeatable they have to
always work on the same data, which is why the test database is
reloaded for each test, either from fixtures or, more often nowadays,
using Factories.
The test environment is not appropriate for performance testing
anyway, to do this you would have to set up a parallel production
environment in order that it correctly mimics your real environment.
On Thursday, July 12, 2012 3:03:25 AM UTC-4, Colin L. wrote:
I followed the guide on creating a test and tried to run it using
something
but I don’t understand how this would scale to a production environment.
Code that works well on 100 or so database rows is not necessarily going
to
work well on millions of rows.
The test environment is designed for testing that stuff works, not for
performance testing.
If this is true, someone should tell that to the official guide
maintainers
The test environment is not appropriate for performance testing
anyway, to do this you would have to set up a parallel production
environment in order that it correctly mimics your real environment.
Exactly! We have this. I just wondered if there was a “rails/ruby way”
of
writing and running performance tests against this environment. The
scaffolding builds a tests/performance directory so I assumed there was.
That munching sound you can hear is me eating my words.
The test environment is not appropriate for performance testing
anyway, to do this you would have to set up a parallel production
environment in order that it correctly mimics your real environment.
Exactly! We have this. I just wondered if there was a “rails/ruby way” of
writing and running performance tests against this environment. The
scaffolding builds a tests/performance directory so I assumed there was.
I see from the guide section 1.7 that for performance test environment
is close to the production environment. Munch munch.