| Commit message (Collapse) | Author | Age | Files | Lines |
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
There was a break to the previous version of the configuration
paramenters API when the configurations parameters were moved to
the database. The actual type of the restart_required parameter
should be a JSON boolean value, and not a string to be compatible
with the previous configuration parameters API.
Change-Id: If45944e25a55813ddf8f80010811da04d667ca19
Closes-Bug: #1379665
|
|/
|
|
|
|
|
|
|
| |
Due to a bad merge, the prepare() call in postgresql does
not have cluster_config argument. Added that to make sure
postgres is in line with other data stores.
Change-Id: I1c3e5032447124dae47ae23ebf820fdcf8080588
Closes-Bug: #1380745
|
|
|
|
|
|
|
|
|
| |
The output of the configuration-parameters-list/show was
inadvertently changed when the parameters were moved into
the database in Juno.
Change-Id: Ic7d215482f5a9ab8ba2059cae7b91a99b35f2795
Closes-Bug: #1379665
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Updated translations by running the following:
$ python setup.py extract_messages
$ python setup.py update_catalog --no-fuzzy-matching \
--ignore-obsolete=true
$ source \
../infra/project-config/jenkins/scripts/common_translation_update.sh
$ setup_loglevel_vars
$ cleanup_po_files trove
Change-Id: I2fcb4e3a0714a73a7026ce2d2c3ffde8aaee1236
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Generates a unique user with a random password for each slave.
The replication user and password is passed to each slave during
the snapshot attach process. Replication user is deleted from
both the master and the slave during the detach process.
Co-Authored-By: Nikhil Manchanda <SlickNik@gmail.com>
Co-Authored-By: Denis Makogon <dmakogon@mirantis.com>
Change-Id: I9cb158a161714bfff90225227f5c652120393ba7
Closes-bug: 1357065
|
|/
|
|
|
|
|
|
|
|
|
| |
This patch adds templates for replica and replica source.
Previously this was hard coded in the code. This gives
deployers the ability to customize their replication
configurations.
Closes-Bug: 1370125
Change-Id: Ie1b4cff1685d2577240eb43ffecac3c977ba964c
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This is a preliminary fix for community review. I expect to be making
some changes as well.
This change addresses some concerns about the Couchbase restore
strategy. In particular, the concerns that are addressed include
string injection in some execute calls, and issues around the
permissions on the secret file used to store the password.
This change set aims to fully address the issue around the password
file but is a partial solution to the string injection
problem. Additional changes are required to completely address
that. Some additional bugs have also been raised to address issues
related to other concerns discovered while fixing this bug. More to
come.
Change-Id: Icd8033caea4866f57f8cb4c5861d21797136ff90
Partial-Bug: #1349939
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Restoring an instance from a backup can take significantly longer than
just creating an instance. Introduce a new timeout for the restore case
with a higher default value
Change-Id: I44af03e9b2c966dd94fcfd0ff475f112ac21be5b
Closes-Bug: 1356645
|
|\ \ \ |
|
| |/ /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
There is no way to tell how long the snapshot for replication
will take, and we have no good way to poll for the slave state.
Eventually, we will need to have an intelligent poll (perhaps
based on guest heartbeats), but in the meantime we will have
the the snapshot use a configurable timeout which can be set
as needed, and independently of the agent_call timeouts.
Co-Authored-By: Nikhil Manchanda <SlickNik@gmail.com>
Change-Id: I6316d748e91d1ec3eebe25a14bb43fbfe10db669
Closes-bug: 1362310
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
If the user attempts to delete an instance which has
replicas attached to it, forbid the delete.
Closes bug: #1362299
Change-Id: I644243aef74d9a34b7726139f9fa5ca8bdb1452c
|
|\ \ \ \
| |/ / /
|/| | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
mongodb 2.6 uses diff config file as opposed to
lower versions. Given diff file candidates code
selects config file that is available
Change-Id: I8fe9eea39a47aa890834a0c55e98c062f9b9fd15
Closes-Bug: #1368468
|
|\ \ \ \
| |_|/ /
|/| | | |
|
| | |/
| |/|
| | |
| | | |
Change-Id: I4a445dedb3a67c0f156f90bb677219d245a78111
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This sync required changes to fix these issues:
* Make execute method clean password in exception
* Make sure mask_password works properly
------------------------------------------------
The sync pulls in the following changes (newest to oldest):
6a60f842 - Mask passwords in exceptions and error messages (SSH)
63c99a0f - Mask passwords in exceptions and error messages
66142c34 - Make strutils.mask_password more secure
-----------------------------------------------
Closes-Bug: 1343604
Closes-Bug: 1345233
SecurityImpact
Change-Id: I3b49b1d667f6ade9ae3f6765d735440a3e838917
|
|\ \ \ \ |
|
| | |_|/
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Many of the Trove configuration options do not have adequate help text
to describe what the configuration option can be used for, and/or what
valid values for the option are.
Since the auto-generated docs are based on this help text, they
currently look very empty and are not very helpful.
example: See "Configure the database" in chapter 4. Database Service
under http://docs.openstack.org/trunk/config-reference/content
All options now have help strings to clarify how/when they can be used.
Also removed periodic_interval option (not referenced in the code).
DocImpact: A separate patch set will be created for the actual doc
changes.
Closes-Bug: #1334465
Change-Id: I3ea1bd255514438c3d356d23d42f3fb3ddd45274
|
|\ \ \ \ |
|
| |/ / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Reasons:
- unit tests are depending on integration tests data, such as:
- datastore and its versions;
- distribution package maintainers are not able to package Trove since ice house RC1.
Gate behaviour:
- gate tests are not failing because unit tests are running right after
fake-mode tests, so sqlite still exist.
Changes:
- fixing those unit tests that are failing while building Trove package
(more info see at bug-report).
Change-Id: Ib9a41d68481d32cdd81fc4d7c015a63c4da33bb3
Closes-Bug: #1302784
|
|\ \ \ \
| |_|_|/
|/| | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The db downgrade scripts are currently not dropping foreign key
constraints, causing errors when the script tries to drop related
tables on downgrade.
This commit address issues in the migration scripts, and also
introduces a new test script to test the migration scripts, so that
issues can be prevented in the future. The new test script is based
on the existing migration test script implementated in Nova.
Change-Id: I240d81afc3e43fd3711de8c156cfb43fd14850bf
Closes-Bug: #1347114
|
|\ \ \ \
| |_|/ /
|/| | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
db_recreate uses repo-path to call db_sync, but for db_sync this
argument is optional so it should be optional for db_recreate as
well.
Closes-Bug: #1279734
Change-Id: I68b0eb3c705525a047eb6bffe8d5e1b4b083cd91
|
|\ \ \ \ |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
It is a pain to load the configuration parameters when setting up
a new environment. Since theres no way to get the uuid of the
datastore version that is created you must get it by looking in
the database.
Closes-Bug: #1371297
Change-Id: If8cea0aaf38878bb04abe6e27b9bfd77d5684bce
|
|\ \ \ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
The code in these three places manipulates the configuration files. I
was (admittedly) having more than the acceptable dose of failures in
execute_with_timeout() and one of them that failed was the move of the
my.cnf file. And this left a turd in /tmp. The failures I was having
with execute_with_timeout relate to other tests that were improperly
mocking/unmocking and some of those problems have now been fixed.
Found that the same kind of thing can happen in cassandra and
mongo. Since default umask is 664, this leaves a file with somewhat
questionable permissions and potentially (at least for MySQL) some
passwords in /tmp.
The files will now be deleted.
Some test cases have been added to exercise the code paths(s) in
mongo, cassandra and mysql. The tests verify that the files are
actually deleted by mocking the os.unlink call. setUp and tearDown
have been updated in each case to make sure that the monkey patching
is properly reversed.
Change-Id: I93302fc5b1b18c9dd7116370a945b94d1824c75a
Closes-Bug: #1354136
|
|\ \ \ \ \ \ |
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
backup_namespace and restore_namespace are now referenced on a
per-datastore basis. Having them in the common options is now no
longer useful. Deleted them from the common_options and marked them as
deprecated in the datastores.
Change-Id: I2d021d68c483b77f9426d78685a06fb0733ec5e9
Closes-Bug: #1351545
|
|/ / / / / /
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
The Trove 'notification_service_id' option has a default value for all
supported databases, except postgresql.
PostgreSQL now has a default value as well.
Change-Id: I033884bc36ba4ffe3faaf6cc4e3496cfb757858b
Closes-Bug: #1373083
|
|\ \ \ \ \ \
| |_|_|_|_|/
|/| | | | | |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
During replication, backup of the master instance is
taken, and a slave is created using this backup.
This patch cleans up the backup which is created
during this process (irrespective of whether it was
successfully provisioned or not).
Closes-Bug: #1359354
Change-Id: Ic636ffeb0bbfb9e6cc02690efe812ffc11bda85c
|
| |_|_|_|/
|/| | | |
| | | | |
| | | | | |
Change-Id: I14ec93eaa97fc52b39e00175adcd718dfc4579b4
|
|\ \ \ \ \
| |_|_|/ /
|/| | | | |
|
| | | | |
| | | | |
| | | | |
| | | | | |
Change-Id: I1b5469ec546694a4f8f48b9500a733c36d316509
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This commit makes the config parameters tests run in fake mode.
Closes-Bug: 1370237
Change-Id: I6ce228a627b35ee418c290a9f19baa163261fb7c
|
|\ \ \ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
A previous commit moved the backup_incremental_strategy option from
DEFAULT group to datastore specific grouping. However, the option was
only implemented for Mysql and Percona so when any datastore that had
backup support (like Couchbase) attempted to start, it received a
NoSuchOptError because of the missing option.
This fix put an empty dict in for the option for all datastores that
don't currently have backup or incremental backup support implemented.
Note: this wasn't picked up by any automated tests because the gate
currently only tests MySQL.
Author: Doug Shelley <doug@parelastic.com>
Co-Authored-By: Amrith Kumar <amrith@tesora.com>
Change-Id: I6a11dcf79c483a43ebdf79085d1bd17d653afe59
Closes-Bug: 1366175
|
| |_|_|_|/
|/| | | |
| | | | |
| | | | | |
Change-Id: I020032ab9a35b0dedfb0875cb717aa89d7abca84
|
|/ / / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Encloses function args in quotes. abc: Foo(foo, bar)
changes to abc : "foo, bar". Intent is to prevent
mongodb json peculiarities from preventing parsing
of the json
Change-Id: Iafd81cf74908078155344ef4746632d48c460f6a
Closes-Bug: #1368373
|
|\ \ \ \ |
|
| |/ / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The test failures were a result of randomized test order and the way
in which these tests were setting up mocks for pexpect.spawn.expect
and pexpect.spawn.match.
The code now mocks pexpect.spawn and then sets return values and match
in each test as appropriate. Code in setUp establishes a patch and a
cleanUp function is added to reverse the patch. No code is required in
tearDown as a result.
execute() was being Mocked using code in setUp and tearDown and this
was replaced that with a similar patch in setUp with a corresponding
cleanUp function to reverse its effects.
Since we're close to Juno, I've not gone down the path of converting
this test to use a fixture which would further reduce duplicated code.
Change-Id: I3edfd499823f2c2f4591e709384a24f1b55abe5a
Closes-Bug: 1365531
|
|\ \ \ \ |
|