This commit is part of a series to retire the Packaging Deb
project. Step 2 is to remove all content from the project
repos, replacing it with a README notification where to find
ongoing work, and how to recover the repo if needed at some
future point (as in
https://docs.openstack.org/infra/manual/drivers.html#retiring-a-project).
Change-Id: I416e08c1d85d78f9c141571bf88b6b7076271088
- Fix all bugs in cleanup:
Removing copy paste, fixing lists, getting id and name methods
Fixing deleting method in case of subresources.
Delete Designate Recrods as well
- Use proper versions of clients in osresources check
- Reduce concurency and amount of iterations in CI
- Remove assertations
Change-Id: I98f61c647051b3f5ce9910d9e613d38450a58b4b
- It's better to provide this paramter explicitly, otherwise for
anybody newbie it'ss going to be hard to understand that
this plugins is related to openstack platform. As well it's
actually part of full name: "<name>@<platform>"
- Fix all places
Change-Id: I15b5ba8b33c542cc3f642dc7940a2eb1677ffefb
In case of workload without data(without results), start_time and
min_duration can be equal to Inf while importing such results via
`rally task import`. Since float column of the table cannot accept
Infinity value in most cases, we need to trasmit a proper value
there.
Also, this patch adds a migration to replca -1 min and max durations
by None.
Change-Id: Ic1a1bf6aac9a994dae4dd0914010c41beec6fbb8
- This change is backward compatible, it will add warning in
case of using deprecated namespace arg
- Use "openstack" platform in openstack context plugins
This is going to be needed after
Change-Id: I76194201aa1f39f3850809f32a0d6b0e55c8f7b1
There is only one module that is tested and it's in rally.common.logging
which means that all tests should be in tests.unit.common.test_logging
Change-Id: I3112f7cb6eb519f167f7ec5302da255b44e90eae
Previously decorates was showing warnings information without
telling what actually cases it.
This commit improves message and adds location of call of
deprecated method or usage of deprecated argument
Change-Id: If6c1ecbdefbd4cc5ed14d3bb938b1f18d078767e
Create a VM, attach N volume to it and list server's
attachemnt.
Measure the "nova volume-attachments" command performance.
Change-Id: Iff7b7aad66b16138beb723ca20d0a7bae7bb24d8
Previously, assertRaise call generates separate output file from each
test function in rally-cli-output-files.
It's better to get the method_name in __init__ to avoid the situation
where some method/call generate new frames. Also I added check when
after getting method_name to make sure rally instance is generated in
each test function instead of being generated once in setup().
Change-Id: Ia7b1608a3abd91caeefbaa021276d7b1a056fa16
os-faults dependency is not lightweight. It depends on ansible which is
quite huge library.
os-faults is not used regurarly, so there is no need to load all its
module and dependencies at run-time for all cases(even when
fault_injection hook plugin is not used).
Change-Id: Iba35853f96ab7d46de12c2fe32da4fd268c90946
This patch changes the output format of plugin information what
is getted via `rally plugin show`.
1.we should not sort the parameters and keep the sort as the same
we commentted in plugins.
2.Remove the redundant blank line, it is ok that just save one blank line.
ORIGIN:
+----------+------------------------------------------------------+
| name | description |
+----------+------------------------------------------------------+
| detailed | determines whether the volume listing should contain |
| | detailed information about all of them |
| | |
| | |
| image | image to be used to create volume |
| | |
| | |
| kwargs | optional args to create a volume |
| | |
| size | volume size (integer, in GB) or |
| | dictionary, must contain two values: |
| | min - minimum size volumes will be created as; |
| | max - maximum size volumes will be created as. |
| | |
| | |
+----------+------------------------------------------------------+
NOW:
+----------+------------------------------------------------------+
| name | description |
+----------+------------------------------------------------------+
| size | volume size (integer, in GB) or |
| | dictionary, must contain two values: |
| | min - minimum size volumes will be created as; |
| | max - maximum size volumes will be created as. |
| | |
| detailed | determines whether the volume listing should contain |
| | detailed information about all of them |
| | |
| image | image to be used to create volume |
| | |
| kwargs | optional args to create a volume |
+----------+------------------------------------------------------+
Change-Id: I10adfaf68d60142e378215136bdf29d3ac597df6
https://www.python.org/dev/peps/pep-0426/#extras-optional-dependencies
Python is a great language and it support extra requirements. In our
case, it can be used for listing database requirements.
This patch adds two separate extra requirements:
* psycopg2 as for Postgres database backend
* PyMySQL as for MySQL database backend
To install Rally with a specific database backend you need just to do
execute pip in right way:
pip install rally[postgres]
or
pip install rally[mysql]
Change-Id: Ibf816758fc093a65b94e6a2b2c273a073275b393
- Simplify rally.common.plugin.plugin code, there is no need to
support from_func() plugins which was removed before, so now
plugins are way simpler to use
- Rename namespace to platform in rally.common.plugin.plugin module
- Switch code to the new plugin interface and fix all tests
Note:
This patch performs minimal changes outside of plugin module
Change-Id: I745539a6b017791d64b42f9303e44027f642b5dc
Glance has an issue with integration with uwsgi. Since OpenStack
community has a global goal - move everyone under uwsgi, glance team
cannot just revert moving under uwsgi, so we need to wait.
Since we do not want to be blocked due to such kind of issues, let's
skip the case when glance resources are not deleted.
Change-Id: I0a8c5d8867f6b43b80a97c2eeda3cefb9f02dd4d
in rally/common/plugin/info replace sphinx prepare_docstring
with trim function from PEP-257 reference implementation
Also move Sphinx from requirements to test-requirements
Added several unittests for PEP-257 trim function
Change-Id: I9f13906cbb3ea9212a78b5bed178e15780001caf
We did not migrate all old atomic to new format, just convert
them while getting data from database and update, so it maybe
occur an error while merging aotmic at line #104. I think it
is the time that we update all old atomic to new format, and
remove the lazy-migration.
NOTE:
It is ok to modify existing migration, since we have 2 cases:
1) the database doesn't include results in old format. In this case the
old migration succedded and change of it doesn't affect anything.
Everything is ok here
2) the database includes results in old format. In this case the old
migration will fail and do not actually apply, so it will be possible
to apply it again(with new code)
Change-Id: Idb2a0ffd161163537e62ced320f8da033f4e04c9
The authentication error message for 'rally deployment check' was
not very specific. I added username and tenant_name information in
debugging mode to make it easier for users to debug. Notice that the
changes are made in credential layer this time.
Also, I refactored the exceptions handling place by concluding
InvalidEndpointsException and HostUnreachableException into one
AuthenticationFailed exception and moving it from
Clients/verified_keystone into Keystone/auth_ref.
Now, verified_keystone will only check if the user is admin and
all other exceptions will be handled in auth_ref().
For tests, I added two unit tests for the Keystone.auth_ref() and
one functional test to check if the authentication error is caught
and if the corresponding details are printed.
Now, the debugging error msg contains the username & tenant name:
"""
[-] Unable to connect for user admin in project demo
"""
Change-Id: I3986f883805e95989c7617cea5396f019a4a0211
There is no need to have two separate classes - users and existing_users.
Both of them are equal from task engine perspective, but require splitting
some logic based on deployment configuration.
This patch simplifies that things and merges "existing_users" context into
"users".
Co-Authored-By: Andrey Kurilin <andr.kurilin@gmail.com>
Change-Id: Ie8fe3b433686bbbd80197876662ccb87b972d5a3
Methods serialize and serialize_data relate to sqlalchemy only and
are used only there, so they should be located there too.
Also, it fixes an issue with serialization of db model object.
Change-Id: I3ceef4d3be3e4f4f95c21ae812992638474684cc
V1 is disable and this scenario fails.
As well v1 is not tested anymore by Cinder team in their gates
so it can cause a lot of troubles for us in future.
Let's just remove it
Change-Id: I387da0dd51922c8a3433e2ef3ed78e254cb20172
* Fix displaying duration of the test
* Use more user-friednly headers in the table
* Add proper functional test for checking the output
* Do not print uuids like a table in uuids-only mode. (it adds spaces
which makes parsing more conplicated)
Change-Id: I8f8f82b6eafd49cee36f4444b681726cce93d6d3
Tony Breeds