As for config options we are using oslo_config library.
It limits us to override config options based on used deployment/env.
To get control of all options and how they are obtained, let's restrict
usage of oslo_config directly and force everyone to use rally.common.cfg
which has the same interface.
Change-Id: Ia845958b50b62d971f1343c03da2552049aa4363
* Move motd to file
* Change borer symbol - old one can 'stuck' some terminals.
* clean pip cache after build
* remove dev-src from sources list
Change-Id: I69df7ba916521920757460e546c066a80017b908
- By default hide creds when running rally db show
- Add option --creds to show them
- Print results of operations as well as connection string to DB
Change-Id: Ifc8ee8ee7b38821409d29ccfe66903e891e12ad3
Now the json format which was exported by `rally task results` is too old,
We expect to use `rally task report --json` instead of it.
New json exporter command:
rally task report --json --uuid <task_uuid> --out <dest_file>
rally task export --uuid <task_uuid> --type json --to <dest_file>
Change-Id: Idbb20513ec1013d0f2b7a7458dc8bffef4771e84
Having 2 entrypoint seems redudant and complicated for end users,
so instead of rally-manage db <commands> let's use rally db <commands>
This reduces amount of code and commands needed to know to work with
Rally which is actually good.
This patch doesn't remove rally-manage just deprecates it
Change-Id: Ib35db2d22174cd3d3eee0d87358ad07a91dce0d9
* Add ability to setup multiple tags for a task.
The database supports an unlimited number of tags per tasks. It would be
nice to support that feature from user side too.
* The tags feature is a bit useless, if there is no way to filter by them.
This patch adds this ability to `rally task list` command.
Change-Id: I959fa904a38f193f77f25b53ba3ee86760fff91c
The OSprofiler is a distributed trace toolkit library.
It helps to trace internal calls of Openstack services including RPC, DB
and WSGI.
This patch integrates OSprofiler in Rally. Rally can trigger the
profiling on a per-iteration basis. To do so a secret key
(profiler_hmac_key) is stored alongside the credentials and used to
initialize the profiler in the constructor of the scenarios. A
configuration parameter (enable_profiler) can disabled the profiling.
Note that in this patch we don't embed the full osprofiler report but
only a trace id. This trace id can be used to retrieve the full trace
from the osprofiler tool later.
Change-Id: I7602856d094e073fde80d287b4d92b5750aacc3c
Co-Authored-By: rcherrueau <Ronan-Alexandre.Cherrueau@inria.fr>
Implements: spec osprofiler
This command could be used to import json results
of a test into rally database.
example:
rally task import --file <json_file> [--deployment <uuid>] [--tag <tag>]
Change-Id: Ie2db6ef19d73ca49e0bf0a4e888f65c07cec3a6c
Default image in devstack was changed recently from uec to qcow2.
All rally jobs except rally-mos now need to be updated accordingly.
devstack changes:
* Id65ebae73b28da7185cb349b714b659af51ef77f
* I5c154ec25555e768954538fc22b4f5d5975b2deb
Change-Id: Ia4d4afb661d8bcc7c38b0a16b508294f789eef14
* Added --filename option to deployment recreate that
will update configuration from provided file.
Change-Id: Icefca03fd86ab8977d0910bf192187c713cb355f
We have a lot of information about verifiers and main horizontal table from
``rally verify list-verifiers`` command could not include all of it.
Co-Authored-By: Yaroslav Lobankov <ylobankov@mirantis.com>
Change-Id: If425e9c373483fa092ebbcefdaf9a46173275c8a
This patch allows us to add some tags to verifications when starting them.
When we have a lot of verifications, it is not a simple task to find the
desired verification from the list. But now we can filter verifications by
tags and it should help us.
Change-Id: I7c6c6a54f31100100b67a750d0381d88fa98dc96
1. Small bugs were fixed.
2. Missing doc strings were added.
3. Now verifier cannot have 'configuring' and 'configured' statuses.
It was done to support work with a few deployments.
Change-Id: I064e863f7597c015e8a8f7c64a4d3049d6baac95
The following changes were done:
recreate -> reconfigure
replace-by -> override
add-options -> extend
Also, some printed messages in CLI interface were improved.
Change-Id: I47bfe88fabc22992b2bcde4c7ea03961c34bc1b6
Two or more plugins can have the same name if they have different plugin
base classes. In the table we did not mention relations to base classes, so
it is hard to identify right plugins.
Change-Id: I31e5633af9a34d59216c281113e32c1839a6172f
Some time ago we added new entity "exporters". For task we created a command
"rally task export" which allows to export task results to external systems.
Exporters can "easily" extende by plugins.
In case of verification component, proper command wasn't created yet. While I
thing idea of exporters is good enough, implementation can be improved.
To simplify usage, entity "exporters" was renamed to "reporters" and
integrated with "rally verify report" command. Generation of regular rally
reports (like html, json) is done in the same way as in plugabble reporters.
Proposed interface:
rally verify report --uuid <uuid-1> [<uuid-2>...<uuid-n>] \
--type <reporter-name> --to <destination>
Examples of usage:
rally verify report --uuid some-uuid --type html --to /path/to/save
# such exporter is not implemented yet, but we expect it soon
rally verify report --uuids some-uuid --type elasticsearch \
--to https://username@passwd:example.com
Change-Id: I4fb38984a73f92503bf2988e509477b10b308cac
There is a possible case when something went wrong and verifier stuck in
configuring state. In such situation there is not way to do un-stuck it,
just full recreation of verifier.
New flag --force should help in such situations.
Change-Id: Icf41f53ec3f6a6a8952a2f5922e1f396ca705ffc
* Split task results into chunks with size
configured by raw_result_chunk_size parameter
* Removed load_finished event as not used anywhere
Change-Id: I09841d02c60ab66eebd851bdf1c0a6d5f9e5a7be
We are planning to rewrite the whole Verification component. It will be done
from scratch. To reduce diff in following patches, let's remove all redundant
"outdated" code now.
Co-Authored-By: Yaroslav Lobankov <ylobankov@mirantis.com>
Change-Id: I64dd3851e381324a8d5274029690b5cfd2333911
In subcommand(for example in verification) we have several groups of
methods(managements, launchers, results). It would be nice to split
these groups in help message.
Before:
compare Deprecated. Use `rally verify results' instead.
detailed Display results table of a verification with detailed errors.
discover Show a list of discovered tests.
genconfig Generate Tempest configuration file.
import Import Tempest tests results into the Rally database.
install Install Tempest.
installplugin Install Tempest plugin.
list List verification runs.
listplugins List all installed Tempest plugins.
reinstall Uninstall Tempest and install again.
results Display results of verifications.
show Display results table of a verification.
showconfig Show Tempest configuration file.
start Start verification (run Tempest tests).
uninstall Remove the deployment's local Tempest installation.
uninstallplugin Uninstall Tempest plugin.
use Set active verification.
After:
genconfig Generate Tempest configuration file.
install Install Tempest.
installplugin Install Tempest plugin.
listplugins List all installed Tempest plugins.
reinstall Uninstall Tempest and install again.
showconfig Show Tempest configuration file.
uninstall Remove the deployment's local Tempest installation.
uninstallplugin Uninstall Tempest plugin.
discover Show a list of discovered tests.
start Start verification (run Tempest tests).
compare Deprecated. Use `rally verify results' instead.
detailed Display results table of a verification with detailed errors.
import-results Import Tempest tests results into the Rally database.
list List verification runs.
results Display results of verifications.
show Display results table of a verification.
use Set active verification.
Also this change transforms all _ to - in cli methods names.
Change-Id: I292e71d159ee35e933119f7fb57209f071aa37d4
* ability to show several verifications results (this
deprecates command `rally verify compare')
* command `rally verify compare' is removed - new report
compares results in better way
* new, unified, AngularJS/Jinja2-based template
* new verifications results processing
* new module rally.ui.report which cares about
report generation (this is a place where code from
rally.task.processing.plot should be also moved)
Examples:
rally verify results\
--uuid <uuid> --html > single_verification_result.html
rally verify results\
--uuid <uuid1> <uuid2> <uuid3> --html > compare_3.html
Co-Authored-By: Oleksandr Savatieiev <osavatieiev@mirantis.com>
Co-Authored-By: Alexander Maretskiy <amaretskiy@mirantis.com>
Change-Id: I942e0d9bf2094f3254dbeccbaa76dbbc3a3ca40e
Magnum renamed baymodels and bays to cluster_templates and clusters
respectively [1].
[1] https://blueprints.launchpad.net/magnum/+spec/rename-bay-to-cluster
Change-Id: I4390406d45a705cf93f8eb6522ebe4a65e468d2c
Partially-Implements: blueprint benchmark-scenarios-for-magnum
Some config options have changed their help text like 'sqlite_db',
so run 'tox -e genconfig' to generate latest config sample file.
Change-Id: Ifad41b2e1203308c02392ad63e3e90853dbbca68
Measure time to create a bay
In this scenario, a single bay is created. This can be used to establish
the baseline in a test environment.
The job yaml file describes the attribute of the bay to be created in
the scenario, such as node count. The baymodel should already be
created by the context: Kubernetes, Swarm, Mesos.
Change-Id: I182abf34212d693d2bb246d195a5daba61e0efe4
Co-Authored-By: Spyros Trigazis <strigazi@gmail.com>
Co-Authored-By: Ton Ngo <ton@us.ibm.com>
Partially-Implements: blueprint benchmark-scenarios-for-magnum
This patch set adds audit template context with two scenarios:
list_audit_templates
create_audit_and_delete
Change-Id: I2d3adef7893fbef1b90e69f56818fbe4d3ce4366
