Kill the old random name generator in favor of the fancy new one.
Implements blueprint: consistent-resource-names
Change-Id: Ia612ebb3ed3af27c6aaa1e35f459b7980b7a24b9
1. The setUp() method was reworked
because no one could understand how it worked.
2. Now ddt module is used where it is possible.
3. Some tests were refactored.
In some tests you will see that
self.assertIn(sorted(expected)[0], sorted(result))
was replaced by
for item in expected:
self.assertIn(item, result)
because sorted(expected)[0] is only the first item from expected tuple.
So we always checked only one item from config section. The wrong behavior
was introduced by the following commit: I487806fe9a8ccd8528922cfaa2700df30961ef9b.
Now it is to be fixed.
Change-Id: Iaabc1a9fbd77e73bd9d28a30b29f9827a22e0bbd
Copied rally/common/objects/endpoint.py to credential.py
Marked objects.Endpoint as deprecated
Modified use of objects.Endpoint to objects.Credential
Variable and key names are not changed in this patch
Partial-Bug: #1447682
Change-Id: Ia62881473d1bf63f952528b66e7420eea6323f08
This switches all wrapper plugins to use the new consistent random
name generator.
Implements blueprint: consistent-resource-names
Change-Id: I6b27fe2f41d84c1961db3b13cf8ddeeae27a6591
Now options 'disable_ssl_certificate_validation' and 'ca_certificates_file'
are configured in the Tempest config file and values for these options are
taken from credentials that user specified.
Config options https_insecure and https_cacert were removed since they were
marked as "deprecated_for_removal".
Change-Id: I74083bc9af21969dd836f852712b81abd3ec74dc
_get_or_create_network: create networks only if there is no networks
in context.
_get_or_create_subnets: create subnets only if there is no subnets
in context.
Unit tests are passing because of bug in ddt [0]
This issue was actually caught by rally-ci (rally-pg-py27-unit) but was
ignored by revievers.
Running tests without testr may be used as workaround:
.tox/py27/bin/python -m unittest discover tests/unit
[0] https://github.com/txels/ddt/issues/35
Change-Id: Ib77bd4f7eeaec2ed3a712e1d56a20d1f87325264
Added support of Swift in Sahara Data Sources. The context does not
verify the output of the job. The cleanup deletes the output container.
Change-Id: I7b4cdc14f6c8a6af227a953be298f266e5851c92
There is no reason to use OOP for one simple method. This patch replaces
class HtmlOutput with simple function "generate_report".
Also, it adds an improvement for skip reason: now ID of a bug is replaced
by the launchpad link to it in HTML report.
Change-Id: Ia75578a67a1fe626b80b877bbda659b5e6070ef0
The current cleanup mechanism for deleting Tempest resources after
the tests finished doesn't work properly. If Tempest tests take more
than 1 hour, we will not be able to delete a resource because of token
expiration. So this mechanism was slightly reworked to fix the issue.
Change-Id: I9807109e6cf7e5652c5fe73c433b7a9a8f3f489d
Tempest creates private network resources for all tests by default.
If a cloud has some shared networks, many Tempest tests will fail with
the following error:
"Multiple possible networks found."
To avoid this issue we have to configure the fixed_network_name option
in the Tempest config file. So this patch proposes the following solution:
if the cloud doesn't have any shared networks, the default behavior
defined in Tempest will be used and Tempest itself will manage network
resources;
if the cloud has some shared networks, we will create our own shared
network and specify its name in the Tempest config file;
if user configures the fixed_network_name option in the tempest.conf
file manually, then we don't need to do anything.
Unit tests for new methods and missing unit tests were added.
Change-Id: I487806fe9a8ccd8528922cfaa2700df30961ef9b
Existing benchmarks which process a list of resources and meters
are updated via new context and adding new configured query
parameters.
It allows to find a weak places in requesting ceilometer
data and extend count of scenarios which use a ceilometer.
Change-Id: Ifbb77915267f7e339ea7adbd9dae34ca59acb78d
*) There is no need to create extra lock object. There is already
instance of lock object in multiprocessing.Value
*) Improve the tests. Unit tests should test that method properly
work and not mock everything to check what is in each python line
(which is anti pattern in unit testing)
Change-Id: I7cc102aae0e083e2fccf98cbc21f8f43cef8d454
subunit2json module was implemented based on subunit2html module, which
cause several issues(see attached bugs for more details).
The main approach of new parser: it designed to be used from Python code,
not as live-filter for a stream.
Since such parser can be used not only for verification, the implementation
is moved to common dir.
Also, this patch:
- removes fake log, which was left from xml parser;
- removes all db calls(except list call) from cli.verify module.
Closes-Bug: #1456810
Closes-Bug: #1471400
Change-Id: Ifcfce827a0df1e75f58e198ee899d4bb85ca9376
This patch adds additional swift scenarios to use what was created
during swift context.
Swift Scenarios:
- list_and_download_objects_in_containers
- list_objects_in_containers
Blueprint benchmark-scenarios-for-swift
Change-Id: I6d615889a1acfcd4686eb061f1041f5b22fd09f0
This permits users to create a network with the network context, or
use the new exiting_network context, to provide networks to Neutron
scenarios instead of always creating a new network. This makes it
possible to run Rally against Neutron Provider Network deployments
with "dumb" routers. It will also eventually permit us to deduplicate
some code by moving network creation into the network context.
Neutron scenarios that are benchmarking network creation continue to
create networks.
This deprecates network creation inside the other scenarios. Warning
messages are produced where appropriate.
Change-Id: I404eee0f14f40e348b2916854ed193c59364c31b