This tests actually were wrong since begging. Cause __lt__ and __gt__
stuff works only in case of instance of classes.
So I really don't know why this worked at all
Add any_order=True to some of assert_has_calls
Closes-Bug: #1358744
Change-Id: Ia699611c14e25b4a1d4592304192c444ba26f853
This adds new resource type that allows to convert VolumeType
name to id before creating load
Remove useless tests as well
Change-Id: I429e261227cf48b5c744836204bbeb972bfdab6a
Function worker_thread is moved out of worker_process cause difficult to
test it when it's nested.
Change-Id: I36d4fa91cfc96769087011f56a848e5987bdba24
Here we add a new "rally info" command, which, based on the input,
prints descriptions for different entities in Rally (in this patch -
benchmark scenarios & scenario groups, more entities are to follow
(deployment engines etc.)). This command makes use of docstrings by
parsing them and printing them in CLI in human-readable form.
The command should be called as "rally info find <query>" (or, in case
of an explicit parameter setting, "rally info find --query=<query>").
Samples:
$ rally info find create_meter_and_get_stats
CeilometerStats.create_meter_and_get_stats (benchmark scenario).
Test creating a meter and fetching its statistics.
Meter is first created and then statistics is fetched for the same
using GET /v2/meters/(meter_name)/statistics.
Parameters:
- name_length: length of generated (random) part of meter name
- kwargs: contains optional arguments to create a meter
$ rally info find Authenticate
Authenticate (benchmark scenario group).
This class should contain authentication mechanism.
For different types of clients like Keystone.
Change-Id: Icf3545c0666d99ab7fd0eaabce8bbe572834e485
This patch adds http_requests scenario which takes a url and an
optional response code and returns the response code of the url
comparing it with expected response
bp http-based-benchmarking
Change-Id: I35373616a266f6bb6410df8afdb23a9229643812
In the semantic validation of the images context it was assumed that the
parameter users is a dictionary that contains the key endpoint, whereas
users is a list of Clients objects. This was leading to a validation
error and any benchmark scenario that had in its configuration file this
context was failing.
Closes-Bug: 1356336
Change-Id: I6452a47975cd3adcd31ec7b7c4ca0159aec8776f
There are several places within the Nova benchark scenarios where we
are assuming networks should be assigned. This breaks a variety of
different deployments. Instead of always assigning a network, do so
based on whether a new scenario argument of auto_assign_nic=True.
Change-Id: I4e2dd58629f3c07b9928a95879056a79382fcf8b
Closes-Bug: #1353622
Keep the scenario_base only on places where required.
Rest keep it as base instead of scenario_base
Change-Id: Ib7887c0a0b65d67a814222e0786fc228cd5bd2ca
Created new directory context/quotas/
Moved to this directory quota.py module
Saperated NovaQuotas, CinderQuotas, NeutronQuotas
in separated files
Similar process done for tests refactoring
https://trello.com/c/ycDv6GD3
Change-Id: I322313d2817a071ab02a386b0fe0d397e3e3cdf0
This magic methods brings a lot of issues:
*) Unclear code
*) Recursion and with construction that makes it hard
to understand traces & as well to have a proper workflow
*) Uncontrollable process of creation and deletion of context:
E.g. it's super hard to "abort" creation of context in the middle
*) create() & cleanup() are inseparable => which makes it impossible
to create "persistance context", that have to run separately
setup() & cleanup()
Change-Id: I071c53900441121767ac3550c5ee5790a358ae73
Add a validator that checks if a benchmark scenario has defined in its
configuration file the context(s) that it requires to run first.
Example:
@validation.required_contexts("context1", "context2")
def benchmark_scenario():
pass
Change-Id: Id83a6ff93acde2abbb29ec8351210e38d725ce48
Pass to validation the full benchmark configuration instead of
extracting parts of it (ex. passing only args).
Change-Id: I8f4cf3e2e8a0392612a0d0828a9c3bd6ab3da90f
* Add flags to output cmd results in json(--json)
and pretty print(--pprint) format
Change-Id: Id4331761e4f9c4bb6735745ed8461eaa671d0945
Closes-Bug: #1341501
- Move atomic actions from benchmark/scenario/utils.py
to benchamrk/scenario/base.py
- Removed the scenario_utils imports
- Updated the scenario_base imports and related changes in
all rally files.
Closes-Bug:#1353464
Change-Id: I7f4d3e4b1d3650d97af08e5aca38fa80cc9f6fac
Recently the neutronclient constructor cacert parameter
has been changed to ca_cert causing rally gate failures.
Fixed it.
Neutron-client Change Id: I9d0395d405b9fbe4db08ad3727f9413be7b82811
Change-Id: Id616d0f315f247d4f89ee779c4424bc93cc208db
Closes-Bug:#1353268
The db model is called 'Worker', it's used to store the rally worker
service status. And the following db apis are added:
* register_worker
* get_worker
* unregister_worker
* update_worker
Change-Id: I38c792cc133dbace10eb7f2d4d202ca05ec7f9cc
This context will be responsable for uploading and tagging an image for
Sahara clusters' benchmarks.
Change-Id: I9b89805e35f48212ac49ec94aa02ffc14d3ab535
One scenraio creates two types of node group templates and executes a
list operation.
The other creates and deletes popular types of node group templates.
Change-Id: Ie37151bb1ff195e6e0a67d85e693333da68f96e9
* Adds "rally verify results --html --output-file output_to_html_file" flag to generate html result from
tempest runs
* json2html generates html report similar to subunit2html using raw json
from rally database
* Use mako templates for html generation
Change-Id: Ib390fa02aba4696bacaa4e4ec8b8c2481a54b0d2
Implements: blueprint tempest-subunit-to-json
'preiodic' runner renamed to rps (runs per second).
Extensible pool of multiprocessing.Process was implemented instead
of using constant length multiprocessing.Pool. That should decrease memory
usage on large iterations and related errors.
Change-Id: I6e367ea30d89d5d0e7ba212ed22e8658421ac9bd
