summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Cleanup the retired repoHEADmasterGhanshyam Mann2020-08-062-0/+4
| | | | | | | | | | | | This repo is not properly retired. Retired repos can only have two files: - README.rst - gitreview To cleanup the retirement, keeping these two files only in right format and removing the other remaining files if there is any. Detail: https://etherpad.opendev.org/p/tc-retirement-cleanup Change-Id: I55f37021fc4fd16d9a175aa361c8503badd80b68
* Retire the trove-integration repositoryAmrith Kumar2017-10-25350-24531/+10
| | | | | | | | | | With the EOL of Newton (impending) and the deletion of the legacy tests[1] and [2], this repository is not required any longer. [1] https://review.openstack.org/#/c/514916/ [2] https://review.openstack.org/#/c/514918/ Change-Id: I8a6466c318029e6c984c6684e92547dd94ae6488
* Show team and repo badges on READMEFlavio Percoco2016-11-301-0/+2
| | | | | | | | | | | | | | | | | | This patch adds the team's and repository's badges to the README file. The motivation behind this is to communicate the project status and features at first glance. For more information about this effort, please read this email thread: http://lists.openstack.org/pipermail/openstack-dev/2016-October/105562.html To see an example of how this would look like check: https://gist.github.com/af3bcb242c9e2828ab264ec09beb108e Change-Id: Ic30c28d25d8fa5fca20fdda8dedf1f92a1e6e2a2 Partial-Bug: 1644851
* Update the Signing Key for Percona Debian and Ubuntu PackagesAmrith Kumar2016-10-244-36/+64
| | | | | | | | | | | | | | | | | | | | | | Attempting to build images fails now with an error because of a change in the signing key for Percona packages. Quick summary copied from [1] Percona .deb packages are signed with a key that uses an algorithm now considered weak. Starting with the next release, Debian and Ubuntu packages are signed with a new key that uses the much stronger SHA-512 algorithm. All future package release will also contain the new algorithm. It’s important that you add the new key before the next release. See also [2] in trove. [1] https://www.percona.com/blog/2016/10/13/new-signing-key-for-percona-debian-and-ubuntu-packages/ [2] https://review.openstack.org/#/c/390342/ Change-Id: I4ca5fccf51e51059d4035f7206609c8afa997c94
* missing yamlAmrith Kumar2016-10-241-0/+1
| | | | | | | Not sure why this just changed but had to fix this in trovestack as well. Change-Id: I3ef0666854df3dfc902a85f876fb0e8881622124
* Change assertTrue(isinstance()) by optimal assertgecong19732016-09-301-1/+1
| | | | | | | | | Some of tests use different method of assertTrue(isinstance(A, B)) or assertEqual(type(A), B). The correct way is to use assertIsInstance(A, B) provided by testtools Change-Id: Ib303f68d83db760dae045f0416f257c4622fa777
* Merge "Enhance 'redstack clean' regarding modules"Jenkins2016-09-271-14/+43
|\
| * Enhance 'redstack clean' regarding modulesPeter Stachowski2016-09-141-14/+43
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Adds a sql command to make sure the modules are not marked as having been applied to an instance so that they can be deleted (otherwise an error may occur if the instance hasn't gone away yet). Enhanced the logic to only sleep if objects are found and to keep trying until the objects are gone (up to 10 times). This makes runs with nothing to delete faster and negates having to run the script multiple times if objects delete slower than expected. Added a check for server-groups to delete them if they exist. Fixed security group delete to look for the 'default' name instead of an id of '1' (which could be incorrect). Also switched it to use the OpenStack client, since Nova was reporting a deprecation warning: WARNING: Command secgroup-delete is deprecated and will be removed after Nova 15.0.0 is released. Use python-neutronclient or python-openstackclient instead. Unfortunately python-openstackclient has a warning of its own: WARNING: openstackclient.common.utils is deprecated and will be removed after Jun 2017. Please use osc_lib.utils Hopefully this warning will go away once the 'openstack' client is updated. Change-Id: If429b2d04111a6e8adbb90c001002215a927d1da
* | Merge "Change MongoDB to version 3.2"Jenkins2016-09-228-66/+167
|\ \
| * | Change MongoDB to version 3.2Matt Van Dijk2016-09-098-66/+167
| | | | | | | | | | | | | | | | | | | | | Cleanup and update the mongodb elements to install MongoDB 3.2 Partial-Bug: 1582712 Change-Id: I7575746ac445a875349640ffa91d6d3fb6680cea
* | | Merge "Replace assertEqual(None, *) with assertIsNone"Jenkins2016-09-142-4/+4
|\ \ \ | |_|/ |/| |
| * | Replace assertEqual(None, *) with assertIsNonejiansong2016-09-112-4/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | From python 2.7 onwards, some new method such as assertIsNone equivalent to the assertEqual,this will simplify the number of parameters. I am looking for all the assert method, which can be updated and renewed,and changed the place where I think they can change. These changes should be harmless refer to [1] for more information [1]:https://docs.python.org/2.7/library/unittest.html#unittest.TestCase.assertIsNone Closes-Bug #1618434 Change-Id: I67db37a0f7561c638d5bb8f2ccd894fa07f6109e
* | | Define credentials before usePeter Stachowski2016-09-121-1/+1
|/ / | | | | | | | | | | | | | | | | | | | | | | In creating the flavors, the variable for credentials is used before it is defined. Although this doesn't seem to affect local use, it causes an error in the gate: ERROR (CommandError): You must provide a username or user ID ... This has been fixed. Change-Id: I5e85c537cf53b1557ca4feaa176484beb3788121
* | Merge "Speed up creation of flavors"Jenkins2016-09-111-7/+10
|\ \ | |/ |/|
| * Speed up creation of flavorsPeter Stachowski2016-09-021-7/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | When creating the flavors necessary for redstack int-tests 'nova flavor-list' is executed for each flavor. This adds over a minute to the kick-start process, even if all the flavors already exists. This change caches the list of flavors and passes it into the add_flavor routine to avoid the overhead of multiple flavor-list calls. Change-Id: Ib70efd4eeb6f6808360d9baaac14edd5a6e84934
* | Add requirements files for newtonAmrith Kumar2016-09-092-0/+61
| | | | | | | | | | | | | | This is to address the upcoming newton branch. Without this, as soon as we branch, newton builds will fail. Change-Id: If270c713dea1900ec6332ba531cdc52a20c2f71d
* | missing requirements of keystoneclientAmrith Kumar2016-09-062-0/+2
|/ | | | | | | The guests are missing a keystone client and therefore won't launch. Change-Id: I102ecfcc169e0cdce93a2c2ad1404d7a929c8152 Closes-Bug: #1620546
* Turn off Neutron until supporting changes landPeter Stachowski2016-08-252-7/+24
| | | | | | | | | | | | | | | | | | | | | | | | There are a number of changesets that need to be merged before Trove can use Neutron successfully. Since two of these are in devstack we'll need to wait to see if they are willing to accept the changes before we can start using it by default. See: https://review.openstack.org/#/c/356026 https://review.openstack.org/#/c/356763 https://review.openstack.org/#/c/356701 Note that the 2nd changeset (devstack) could be rolled into the Trove plugin (3rd changeset) by copying the relevant functions (not ideal, but doable) however the 1st changeset (devstack) is fairly critical and would be difficult to implement outside the devstack Neutron l3 plugin. Also changed the logic for the Neutron client (and others) such that the code is pulled from git if the project is enabled, and made sure *_BRANCH and PROJECT_BRANCH were honoured on the plugins. Change-Id: I764f15514a40926173b6c9194ddbf24c6c1f999a
* Use Neutron by defaultMatt Van Dijk2016-08-188-39/+37
| | | | | | | | | nova-network is being deprecated, and devstack is moving to using Neutron by default. This patch makes redstack use Neutron by default and fixed a number of issues related to using Neutron. Change-Id: If216688ef26b118c6d7598b2e086ba446b03978a Closes-bug: 1604019
* Merge "Add "clean" command to redstack"Jenkins2016-08-141-0/+47
|\
| * Add "clean" command to redstackDoug Shelley2016-08-091-0/+47
| | | | | | | | | | | | | | | | | | | | | | | | A long time request has been to add a "redstack clean" command that would clean out all the trove, nova, cinder etc resources left around after a failed test run. This commit adds a "clean" command to redstack. It is a first cut and hopefully can be evolved by the community over time. Change-Id: I5e2e28936161e65e0a4edfa5c4b9d5ffcab6cee8
* | upper-constraints.txt handling in trove-integrationAmrith Kumar2016-08-094-20/+60
|/ | | | | | | | | | | | | | | | | | | Currently, trove-integration attempts to get an upper-constraints file into the guest by looking at the tip of the appropriate branch. This is fine for developer side testing but not ideal for the CI. This change looks to see whether the CI setting (DEST) is set, and use that to see whether $DEST/requirements/upper-constraints.txt is a file, and if yes to both uses that. If not, it attempts to guess based on the branch name. This therefore covers both cases, the CI and the developer environment and should allow for chaining with jobs that attempt to change u-c. Co-Authored-By: Tony Breeds <tony@bakeyournoodle.com> Co-Authored-By: Amrith Kumar <amrith@tesora.com> Closes-Bug: #1610938 Change-Id: I06406051c849c924371f22b9fe1b03ab99e300f2
* Merge "rename a file for fedora guests"Jenkins2016-07-191-0/+0
|\
| * rename a file for fedora guestsAmrith Kumar2016-07-101-0/+0
| | | | | | | | | | | | | | | | fedora guests don't use upstart. renaming the file because the current name (20-guest-upstart) while identical to the name for the ubuntu file is confusing and misleading. Change-Id: I9818235e0d367c8213ccd1f60c4b26be11dd3735
* | [fix gate] Use upper-constraints on image buildPeter Stachowski2016-07-154-2/+38
|/ | | | | | | | | | | | Currently building an image uses requirements but not upper-constraints. This can cause an issue when libraries change and the image becomes out of sync with the server code (for example when oslo.context changes). Trove integration now uses upper-constraints as well when building images. Change-Id: Iba5e9976ff9522b7b6fdd523ed78fc6d9b70ea6c Closes-Bug: #1603506
* Install PyMySQL on guestVictor Stinner2016-06-282-0/+2
| | | | | | | | | | | | Add PyMySQL (>=0.6.2) dependency. PyMySQL is now the preferred MySQL client, for example it supports Python 3. See the Trove change I65e8a8d5dc251a8b00529cdfb1a6ada3d5720f68 for the full rationale. Use pip dependency rather than the python-pymysql package, because the package was only added in Ubuntu Wily (15.10) and Fedora 22. Partially implements: blueprint trove-python3 Change-Id: Id4d013d174ba40a453819f900aaa316a93e59b48
* Merge "Have Trove plugin install dashboard"Jenkins2016-06-252-6/+7
|\
| * Have Trove plugin install dashboardPeter Stachowski2016-06-172-6/+7
| | | | | | | | | | | | | | | | | | | | | | This fixes redstack since the Trove plugin now handles installing the dashboard if Horizon is enabled. Variables were changed to reflect current names. Change-Id: I18f68136ebb81bc4b4f8c12062d084a5d8901193 Depends-On: I6586801bff4282e3bbf11e45e4f171fffcab285e Closes-Bug: 1593806
* | Merge "Move the rabbit/rpc options to its own section"Jenkins2016-06-241-0/+1
|\ \
| * | Move the rabbit/rpc options to its own sectionAmrith Kumar2016-06-211-0/+1
| |/ | | | | | | | | | | | | | | Using the DEFAULT section for rabbit configuration is deprecated, so all rabbit related option should be moved to the oslo_messaging_rabbit section. Change-Id: Iec072877e9174e136b4f0c1c806d7f9b377a7f2e
* | Merge "Add another 'fault' flavor"Jenkins2016-06-232-4/+8
|\ \
| * | Add another 'fault' flavorPeter Stachowski2016-05-162-4/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Apparently there are two code paths for a failure creating a Nova instance. One returns immediatesly and the other takes some time to fail. The case where the initial command to Nova succeeds only to fail later needs another flavor to test it. A new flavor with not enough disk to install the Trove image has been created to test this scenarion. Change-Id: I5d3339e9cbfd6aeb0c3ff6936fefa8dbe9e841f8
* | | Support running Mistral under RedstackMorgan Jones2016-06-082-0/+11
| |/ |/| | | | | | | | | | | Adds support for ENABLE_MISTRAL option to configure Mistral to run under redstack. Change-Id: Icddd5b34c30c5774e6b454c24f8373d30ca2da22
* | Merge "Trove image build fails getting percona key"Jenkins2016-05-184-0/+60
|\ \ | |/ |/|
| * Trove image build fails getting percona keyAmrith Kumar2016-05-064-0/+60
| | | | | | | | | | | | | | | | | | | | | | | | | | This fails often in the gate and I ran into it migrating trove-integration into trove-image-builder. The most common cause for this is that keys.gnupg.net fails. While there are multiple servers proxied for this service, sometimes they fail. In general people have found keyserver.ubuntu.com to be more reliable. Change-Id: I97dcb9120c75129d8a0aaed9ddd39b9e5fa7e608 Closes-Bug: #1579094
* | Add 'fault' flavor to test persisting an error msgPeter Stachowski2016-05-113-4/+9
|/ | | | | | | | | | | | | | In order to test that error messages are now persisted, a method of guaranteeing that a Trove create command will fail is needed. The simplest way to do that is to attempt to allocate more resources from Nova than it has available. To this effect, a 'fault' flavor has been added. Also bumped up the disk on mariadb to handle more operating systems. Change-Id: I32a5ed229988f047db44183c71f0462fa6f1990d
* Merge "Change redstack to use default config values"Jenkins2016-05-031-9/+12
|\
| * Change redstack to use default config valuesPeter Stachowski2016-04-051-9/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The plugin code has been enhanced to allow configuration of Trove using variables. The default there is to use the actual config values in the Trove repsoitory. This has been mirrored here, in that the variables are listed so that a user/developer can modify them if they wish, but they do not override the default values anymore. The proper place to add these variables would be in a 'local' options file, such as options.rc or ~/.redstack.options.rc Change-Id: I38bc7190f1c112b4ae3cd6614f1d6cb625b3387d Depends-On: I2c568fdfa05064682f372b996a5aebcdd4e93ef3
* | Merge "Add xmltodict to kick-start requirements"Jenkins2016-04-232-0/+2
|\ \
| * | Add xmltodict to kick-start requirementsPetr Malik2016-03-072-0/+2
| | | | | | | | | | | | | | | | | | | | | | | | Add a requirement for: Ia57587bf557771ea6d7f00e2fe9d674789b98559 Change-Id: I254b81b45e44aeda3049865bee76cd9075312761 Related-Bug: 1550557 Depends-On: I4a088b7a541c35a6c32d6985727bad65ff491815
* | | Merge "Fix "trove_volume_size" test prop type"Jenkins2016-04-2213-74/+96
|\ \ \
| * | | Fix "trove_volume_size" test prop typePetr Malik2016-04-1413-75/+97
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 'trove_volume_size' test property should be an integer. Mongodb and Postgresql test configs define it as a string causing the scenario tests to fail on these datastores. Also add a large-disk flavors for datastores that require more space on the root volume (Vertica). Change-Id: I352580bc4967ba2d7125e26f037ca4bcda31b2b4
* | | | Merge "Don't reset HOME in redstack"Jenkins2016-04-171-11/+3
|\ \ \ \
| * | | | Don't reset HOME in redstackAmrith Kumar2016-04-171-11/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently redstack resets HOME as the first thing that it does. It isn't particlularly apparent why this is required; it may have been needed in a long ago time when paths to thing were trove-integrations/scripts relative but they aren't now. Change-Id: If0f7329ea37f6d92f10b815c7a96eaee97d8b6d4 Closes-Bug: #1552048
* | | | | Merge "Elements for PostgreSQL 9.4 and pg_rewind"Jenkins2016-04-165-14/+100
|\ \ \ \ \
| * | | | | Elements for PostgreSQL 9.4 and pg_rewindAlex Tomic2016-03-045-14/+100
| | |_|/ / | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Failover for postgresql requires pg_rewind, which needs to be compiled for 9.4 Change-Id: I226955dd8af1107b63685bf71e37859eb924d487 Partially-implements: bp/pgsql-replication
* | | | | Merge "redstack repl is broken"Jenkins2016-04-161-15/+2
|\ \ \ \ \ | |_|_|_|/ |/| | | |
| * | | | redstack repl is brokenCraig Vyvial2016-02-011-15/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | remove the nova imports to get the repl to work. Change-Id: Ie919b7ac1c95a2b69e7800cb2837ce0123e48690
* | | | | Merge "Allow redstack to force kvm in devstack"Jenkins2016-04-043-1/+6
|\ \ \ \ \
| * | | | | Allow redstack to force kvm in devstackCraig Vyvial2016-02-013-1/+6
| |/ / / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Force kvm is useful if you have an environment where you can use kvm as the virtualization for nested instances. Change-Id: I89ac30243e3c47cbd36b0a0890242a2a8031bea4
View Lorry Controller Status