| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
Change-Id: I8f943f90df8cd8dd0d322c2305f4dfa2d0521c81
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Global name '_LE' was missing in trove/cmd/guest.py. This patch
adds this missing import.
Change-Id: Ia0b4ecb27f0f4dcae88a56288bc5052f1afa63e8
Closes-Bug: #1620888
(cherry picked from commit e428d76ace2e21d83991243ad00b4951ae05da6a)
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
There is a test to eject a valid master during replication that is
failing more and more often in the scenario tests. Basically the
eject will only work if the heartbeat from the current master is
more than a minute old. This shouldn't happen in the scenario
test run, but it does - quite often.
Since this is consuming a large amount of gate resources, and the
bug isn't that onerous (but is probably hard to fix), the current
'eject valid master' test will be turned off until the fix lands.
The scenario test will print out the bug number during each run
as a reminder (using the new SkipKnownBug method).
Co-Authored-By: Peter Stachowski <peter@tesora.com>
Co-Authored-By: Amrith Kumar <amrith@tesora.com>
Author: Peter Stachowski <peter@tesora.com>
Change-Id: Ia543da551ad4394d4964541f9db474e0792b9337
Related-Bug: #1622014
(cherry picked from commit 7d8d743d8e168f64fa893c975d7301507d4985d8)
Conflicts:
trove/tests/scenario/runners/replication_runners.py
trove/tests/scenario/runners/test_runners.py
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This changes the default setting for use_nova_server_config_drive
from False to True. By default, nova does not use a config
drive for launched instances. Therefore trove must implicitly
request the config drive be used so that it can inject
guest_info.conf into the guest VM.
This commit also adds an error to guestagent when guest_id
is missing. If the guest_id configuration parameter is not
set at guestagent startup it will throw a RuntimeError. This
case typically occurs because guest_info.conf was
not injected into the guest, or was not included in the set
of configs that guestagent reads at startup.
And finally, this commit adds a section in the 'building guest
images' documentation describing how configuration is injected
into the guest agent.
Change-Id: I084c841472183893a63ca7b70d560f44a2f08901
Partial-Bug: 1609915
(cherry picked from commit f7b8805c3aed5b5cef01176cc1d9136bd31125fb)
|
|
|
|
| |
Change-Id: Ib74c92123a9acc8e39fa9caec49d929ba7265628
|
|
|
|
| |
Change-Id: Ia67c1ae4c0c3ea0b08a01b357ebbb09c2110b2d2
|
|
|
|
| |
Change-Id: I642e789758617819b7f69739618afde796bac10c
|
|
|
|
| |
Change-Id: I8c301761cb0a3bb4261f689d9dc5d81f7fa29f8f
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
In MySQL 5.6/MariaDB 10.1 and later versions there was a change
on the restrictions over foreign keys (FK).
In 019-datastore-fix.py, we attempt to change a column that is part
of a FK from NULL to not NULL. This sort of modifications are not
longer allowed.
This situation caused trove-manage db_sync and trove-manage db_migrate
to fail.
To workaround this, the FK check before executing the ALTER query
is disabled and then reenabled.
Change-Id: I666d01235f2c3225aca3fe7520ebdf6d53831cab
Closes-Bug: #1473226
(cherry picked from commit 679e2283ecc888bd567955b4bf431c87043e8d12)
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Adding a helper client for MySQL will enable int-tests
involving actual data operations.
The test_helper.py file was also modified to push
more boiler-plate code into the base class, as it has
become apparent that only one 'add' 'remove' and 'verify'
method is probably adecuate for most datastores (given
the right sane parameters). The Redis helper was also
cleaned up to use the new format.
Change-Id: Iff34ca0b922adc9eea8b3861b05adf0251109221
(cherry picked from commit 2b98c538624a55bda9abe304c08b19435869cbc6)
Co-Authored-By: Petr Malik <pmalik@tesora.com>
Co-Authored-By: Peter Stachowski <peter@tesora.com>
|
|\ \ \
| |/ /
| | /
| |/
|/| |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Add generic integration tests for instance create.
Also moved some global constants to the base runner.
Note: Did some additional cleanup work on other test groups as well.
Broke down some negative tests into more self-contained units.
Added a comment string on tests that did not have it.
Had to fix database_actions test to work for users defining
hosts (like MySQL users).
Change-Id: I7c1413fa5742c276454e26ebbbeae8e02569a392
(cherry picked from commit aeadbfc974fc515ef85059c48277a56d0ec9037e)
|
|\ \
| | |
| | |
| | | |
stable/liberty
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This will force pip install to use the upper-constraints.txt specified
version of pip modules. When you don't do this, you are out on the
bleeding edge and become unstable everytime some python library in the
wide world changes in a way that you don't expect.
The releasenotes job has no constraints file, so override the install
command there.
This commit is being proposed on stable/liberty and therefore the
install command includes the clear designation of the branch. This is
because, on branches other than master, we need to be sure to get the
upper-constraints.txt file for that branch.
Change-Id: I9c37be55769d5a4a9ed1704f36302596a5f9ee2d
Closes-Bug: #1563038
(cherry picked from commit aca90168f52169afb36500d6af77723b1a707008)
|
|/ /
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Routes v2.3 which just came out included a change (see the LP bug
referenced below) that caused py27 to fail.
At issue is that Routes<2.3 allowed both
submapper.connect(route_name, route_path), and
submapper.connect(route_path).
In 2.3, you were required to provide both route_name and
route_path. This change adds a route_name to support both Routes < 2.3
and Routes == 2.3.
Change-Id: I8738d4dcac2cc37b727767ba614fab4339758e22
Closes-Bug: 1563028
(cherry picked from commit 6664fb27af2218c411230f91b8e23e72845cbf2a)
|
|\ \ |
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| | |
It turns out that some modules may throw exceptions
other than ImportError from their class members.
The mock detector should ignore those.
Change-Id: I513c573982250e24b94ec5e649cf2c9e2799334c
Closes-Bug: 1524918
(cherry picked from commit e8d9a3647acc4841f401cb650c5649e87b0a8d6f)
|
|\ \ |
|
| |/
| |
| |
| | |
Change-Id: I48e61a212f904171fd7128c3e705c32190464ceb
|
|/
|
|
| |
Change-Id: Ia0520aff6d96ce3fcf20e94e2a100dae0f90aa76
|
|
|
|
|
|
|
| |
Update .gitreview for stable/liberty branch to simplify submitting
backports.
Depends-On: If72ea960f6479e73234ec0cf1af10687f172b86b
Change-Id: I394ecd1a98ce286b49bbacb62674a3c80182ec51
|
|
|
|
|
|
|
|
|
|
|
|
| |
No longer mention next version on setup.cfg on stable branch.
oslo.utils 2.6.0 causing tox test to hang.
This forces a specific version of oslo.utils *before* the
requirements are loaded, thus ensuring that 2.6.0 is *not* used.
(This is a temporary fix until oslo.util pushes a new version.)
Depends-On: Ica981756c67902e7fd7bd07ff76c5b679e8e4e2f
Change-Id: I97ec9d17b6a611d888ef007575a29c1811f475c9
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The 'set' command in the Python client interpretes None as the
string 'None.' If no password is set on the master, a detach/attach
(such as is done with promote) will have the slaves try and use 'none'
as a password. This will cause them to not be able to connect to the
master and leave them in a detached state.
The 'masterpass' is now set to an empty string if no password is set.
The restart of Redis was also moved before attaching the slave,
so that the mapped CONFIG command is recognized properly. This is
required so that setting the master password completes successfully.
(cherry picked from commit 6c4db5d327df7e10f7a0bf38deaee0db4fe2ee5a)
Closes-Bug: #1498208
Change-Id: Ibc842b95fd191908d4fac420030f84393c696013
Depends-On: Ida0aa155b1e1ec7b8e9fa73006d3db9a88ffea02
|
|\ \ |
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The nova client created by create_nova_admin_client was throwing
an Authentication exception. In order to properly create this client,
novaclient needs tenant_id set to a tenant-id. Note: novaclient
documents that project-id is specifically a tenant-name and from
what I can tell it isn't required for Trove use-cases.
Added a unit test to check that create_nova_admin_client causes
a nova client to get instantiated with the correct params.
(cherry picked from commit e1538ae75e37b14905893d91c45dd044d32a7eb0)
Change-Id: I34d7d9dd36117bff796db22b851656ba78988558
Closes-bug: 1473550
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In preparation for the release, do some cleanups for translations.
Removes all po files that are partially translated. The translation
team has decided to exclude files with less than 66 % of translated
content. There is no content lost, all data is in the translation
server, we just remove it from the repository (not applicable to trove).
This updates also recreates pot (translation source files) to
reflect the state of the repository in case there was no recent
import for them.
This change needs to be done manually since the automatic import does
not handle some of these cases.
Change-Id: Iefe0a9ce6f980fdc1f107e9bb25a171fd505b135
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Currently pxc only allows for creation of a cluster having exactly
3 instances. Fix this so that we are able to create a cluster
having more than 3 instance members.
Change-Id: I4f85ccc00b95dcded8adc01dcf8a3ad7aabaf792
Closes-Bug: 1494584
|
|\ \
| |/
|/| |
|
| |
| |
| |
| |
| |
| |
| |
| | |
Added a new remote.py using which deployers are able to configure
Trove to deploy instances into a single tenant.
Closes-bug: 1498592
Change-Id: Iadaea2623cf1962a0b8a078e2a702c3aea85c176
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Use the existing shared operating system interfaces to manage the
Postgres service.
Use the 'state_change_wait_time' configuration value
instead of the currently hardcoded '30s' as start/stop/restart timeouts.
The existing postgres implementation does not wait for the
database to change its status after a service call.
This behavior can potentially lead to severe transient issues,
such as FS corruption (unmounting the volume when the database is
still writing to it) etc.
This patch fixes the above issue by making sure to always wait for the
database to actually start running or shutdown before proceeding.
Abstract the start/stop/restart database calls to tested reusable
interfaces of the Status class, so that all datastores
can make use of them rather than having to re-implement and test this
common functionality from scratch.
It should hopefully help to cleanup guest-agent implementations and
prevent the above issue in the future.
NOTE: The added calls would be good candidates for base-class
methods once we have a base service class for the guest-agent.
Closes-Bug: 1492526
Change-Id: I0ab198be4f96e3ba3c2e8a158d11de3a44046ab3
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The test in mgmt.test_datastores.TestDatastoreVersion.test_version_index
is failing because of concurrency race problems.
Executing in sequence the delete method would delete one
of the datastore versions and
while executing another test for index method,
the setup method would recreate the deleted object
so that queries work correctly.
But because of delete and index methods running concurrently
on different threads, the delete is deleting the object
during the execution of setup/index method on another thread
because of which the object query is failing.
As a fix, did two things,
1. Since 'test_vr1' is only used by delete method,
shifted the query of test_vr1 to delete method.
2. Updated the mocked method used in test_version_index from
models.Datastore.load to models.DatastoreVersion.load_by_uuid
the former is not needed in test_version_index,
the not-useful mock got left, probably,
while the code for index method was updated.
Change-Id: I816b602579892588d3179a0dc08b9fc5668f1e5b
Closes-Bug: #1492130
|
| | | |
| | | |
| | | |
| | | | |
Change-Id: I1acc2257344304bb6cba7242aac48b16362ed87b
|
|\ \ \ \ |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
report_root is being called in the case of a new instance being
created from backup when root is enabled on the originating instance.
A change [1] was made that added the context object to the report_root
argument list and the Mongo guest wasn't passing it.
[1] https://review.openstack.org/#/c/124355/
Change-Id: Id61337ae85fb980752d606bddb2f1e574a345cba
Closes-bug: 1498568
|
|/ / / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
During cluster creation, we are now able to pass in custom
properties specific to a datastore as part of a data-bag
named 'extended_properties'. This allows us to get these
properties to the appropriate clustering strategy which can
then use them to implement datastore specific actions.
Change-Id: If24f5bf6181bf1502042ac51868a161a6a037bf2
|
|\ \ \ \
| |_|/ /
|/| | | |
|
| | | |
| | | |
| | | |
| | | |
| | | | |
Change-Id: I2931badfe53221e78329c78d93f27cd605d2a6cc
Closes-Bug: #1487711
|
|\ \ \ \ |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
The get_flavors unit test sometimes fails as the list can
come back in a different order. This causes the json comparison
to inacurately claim that the content is different. Since
we don't care what order the json data is stored, this
comparison is now done on a sorted data set.
Closes-Bug: #1437179
Change-Id: I92eaa1c98f5a58ce124210f2b6a2136dfc573a29
|
|\ \ \ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Change the use of "nogroup" on a chown command to use
MONGO_USER instead.
Change-Id: Ic605733188830dd085a597ecf9e8669ad6247f83
Closes-bug: 1487903
|
| |/ / / /
|/| | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Since most of us don't have MongoDB installed on our
desktop/development environment, tox tests will fail because some
config paths for Mongo aren't available.
The change proposed here is merely to mock the _init_overrides_dir
properly in changes that would otherwise stub their toes on this.
It is my understanding that this issue was introduced in Change-Id:
Iaca9cc9e3ce1e6521839062fc9458c1b987f9464 which sought to address bug
1485158.
The reason this problem occurs on my machine (and likely on other
developers workstation) is that I don't have MongoDB installed. The
reason this doesn't impact the gate is that those machines have
MongoDB installed (for ceiliometer).
Change-Id: I0f3d24b95a7d2ef1e23a0f1fd8f60bc496261f70
Closes-Bug: #1495573
|
|\ \ \ \ \ |
|