Create Makefile and recreate conf.py using sphinx-quickstart
Add _build to gitignore
Fix several nits in rst files
Add tox environment for docs building. Now docs may be built by
running `tox -edoc'
Change-Id: Ib0eb17047e583bdf3ac9111e117f908340f4d5ef
Adds "scenario_output" and "atomic_actions" keys to constant and
periodic runner result when exception is raised while running a scenario
Change-Id: I5ddc8d08e787eab34d67df400acb33a133ea0cef
Closes-Bug: #1325630
This patch contains next main changes:
- two scenarios for heat project ("create_and_list_stack",
"create_and_delete_stack").
- three atomic action for heat project ("stack-list" , "stack-create",
"stack-delete").
- two task samples how to use new scenarios.
- a little changes in utils for supporting heat stack_status property.
Change-Id: Ice2bd4dc2fcefb7351df69dd8463abc6f8d4fd24
* Log deployment id and benchmark key from task to help when
reading logs from multiple benchmarks which are concatenated
together
Change-Id: Ia5690cfd911c38f71c3ed4595cae5f2aa683f842
* added types.set(...) decorator to apply transform to scenario args
* abstract ResourceType implements generic transform logic
* support use of flavor name through FlavorResourceType
* support use of image name through ImageResourceType
* transform is applied during validation and during preprocessing
which follows validation
* modified all sample scenarios to use flavor-by-name and
image-by-name
* passing resource as simple key-value pair will no longer work,
and instead resource must be object containig id, name, or regex
* unit tests for image and flavor transformations
Change-Id: If35959ece8f247b5376f5ff91ae3d3e008a84447
* When uploading an image to glance from a file,
perform the close operation from finally block
to avoid waiting for handle to be closed when
reference count goes to 0
Change-Id: I5c3ee35b87fa55097c2e847836d5ad1764ae0574
Without scenario name users are not able just to copy/paste & run
the scenario, this patch add the missing name for the configuration.
Change-Id: I4b5f8cd1b6f0d03e264580c84bb06b7bada4792a
The benchmark scenarios are not sorted in the selection field
in the output html. This patch fixes the problem.
Change-Id: I11bf545f659dccd088b1ff16961ae89be3aee8dd
The CLI code has recently been partially rewritten,
so that it uses now the openstack.common.cliutils.print_list()
method to draw tables, instead of pure PrettyTable library usage.
Change-Id: I895112c5514a1f9e31919e110d16650272ca5960
The sed command (line 138) responsible for setting the database
connection string in the configuration file of Rally, is only GNU sed
compatible. The BSD systems (ex. FreeBSD, OS X) use a different sed
implementation. The in-place editing differs in these two versions. This
commit modifies the sed command in a way to be compatible for both
implementations.
It creates a temporary file and sends the output of sed to that file,
and then it replaces the original file with the temporary one. It is
actually what the --in-place switch of GNU sed does in the background
[1].
[1] http://www.gnu.org/software/sed/manual/sed.html
Change-Id: Idef71121416c1c6c5d70ab6e041f89a7b441b1f9
Fix several issues in rally.deploy.engine.EngineFactory and
rally.deploy.serverprovider.provider.ProviderFactory:
* Wrong API descriptions
* Wrong module references
* Grammar
Change-Id: Idddf3da0c935ba7d21132c67be9259a1fbf96130
This would simplify the way measure atomic opeation duration
in certain cases. For example if we want to get the duration
for each operation which runs in an iteration
for i in range(repetitions):
with scenario_utils.AtomicAction(instance_of_base_scenario_subclass,
"name_of_action"):
self.clients(<client>).<operation>
Change-Id: I48793fc0af160656f37964ca5d4e9b59b62de685
BaseException.message has been deprecated as of Python 2.6.
http://legacy.python.org/dev/peps/pep-0352/
Replaced e.message with six.text_type(e) (uses unicode with
Python 2.x and str with Python 3.x).
Change-Id: I8d0c43f9383cec343ca159061b7a93d3db45a74a
This patch replaces redundant code which generates random names
with calling classmethod base.Scenario._generate_random_name,
for all benchmark scenarios.
Change-Id: I051f83508e580f831ae665e09ee20275b4c77ded
The stacked and pie graphs can still be produced when there are no
successful iterations. The pie chart ilustrates this well by showing 100%
error. However the histogram cannot be produced as the number of bins
cannot be calculated, so this is skipped.
Also hides atomic actions header and select when there are no atomic
actions results.
closes-bug: #1315063
Change-Id: Ic1fb7f00550828a8e374a94034c1b8eb07db5e89
This secnario will throw an exception with the probability provided.
Useful for testing where some good iterations and some bad iterations
are required.
Also adds:
* number validator to dummy_exception scenario.
* more itterations to dummy scenario sample files
Change-Id: I50360de0347a8c28686a3196edefe3ae004fce85
