Retrieve the user names and access rights
in a single query hence closing the race
window.
Change-Id: I66847827fa8cb0615c1662670017d2cc55466ff3
Closes-Bug: 1617464
Support for standard WAL based streaming replication for
postgresql guests. Sets up read-only hot standby servers.
Basic replication is supported for 9.3, but promote
requires pg_rewind which is only supported for 9.4 and
greater. Promote currently only works with single
master-replica configurations (see TODO and associated
bug 1553339 for details)
Test changes:
- moved the replication promote tests to a separate group
- skipped known issues with bug reference
Change-Id: I519fa58d786edddb09bf14ce7629e1be51b62600
Implements: bp/pgsql-replication
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
The add_members call was using a hardcoded value when there is a
configuration variable that should be used.
Change-Id: I921ebddb581a52f92e427a8bdcbbb31650846125
Closes-Bug: 1622028
Recently we added an os_admin user to PostgreSQL for Trove use.
PostgreSQL requires a database for each user, so this now appears in the
database-list command. It needs to be added to the 'ignore_dbs' config
option so that this doesn't happen.
The tests don't capture this since they use the cfg setting to make sure
that no 'unwanted' db names are returned - a bit of a chicken/egg
problem.
Change-Id: Ia900826d9f2d775466dabde433e9ac1349148112
Closes-Bug: #1621997
This change updates the api-ref documentation with some issues
identified in an earlier review.
Change-Id: Ic769a1ae376e8569f5a13af8feada88f4cf0ac32
Closes-Bug: 1614923
Related: I3315261aa18729fa7a6aa79d4a1d6c24de1e2c6b
Old installation instructions are outdated, will happen
"HTTP request sent, awaiting response ... 404 Not Found
2016-09-09 02:21:09 ERROR 404:. Not Found "
I think it should update instructions, and I think
the way is to let the user to download on this page,
because the relationship between md5 can not generate a fixed format
Change-Id: I1baf30a59969b003abb044e12617609f2d213a39
The scenario tests for clusters are occasionally failing when doing a
cluster-show after a cluster-grow. This seems to happen because the
locality flag on the new instances takes a few seconds (sometimes) to
populate. This results in a 'AssertionError: Cluster missing attribute
locality.' error in the cluster grow test.
The fix is to check all the instances for the server_group information
instead of just the first instance.
The cluster tests were also tweaked so they could be run separately.
Change-Id: I4f369231a8f0bc8cad338e2eb1de9cd1ca0253f6
Closes-Bug: #1621634
In order to support automatically updating the release notes when we
create stable branches, we want the pages to be in a standard order.
This patch updates the order to be reverse chronological, so the most
recent notes appear at the top.
Change-Id: Ib364dcc8eb31275a31c83b68d7914263b183e393
Signed-off-by: Doug Hellmann <doug@doughellmann.com>
At the time of test setup, repeat the definition of the several
variables, and does not change the contents of the variable in a
method, they can be declared as global variables to the to simplify
the setup
Close-Bug: #1621464
Change-Id: I9c00c17e85c591671e0a104c004d6465b7bbb72a
The code reads
else:
print("Starting server:%s" % pid)
pid_file = CONF.pid_file
with open(pid_file, 'w') as f:
CONF is not defined
I tried to add this information to the common/cfg,
but I'm not sure this add is correct, please help
me to check
Closes-Bug:#1617973
Change-Id: I31e63b0321da912f6b59fbbc901e0141a7d1e73f
I deleted FakeFlavors # self inside _add (13, 20, "m1. Heat", 512),
nova inside have no this configuration of the test, in fact this
is not in our test start, so I think this test item is useless.
some consult:https://github.com/openstack/nova/blob/master/nova/tests/fixtures.py#L315
Change-Id: I6afe2492b157c580efc15099d76281ba8ebe6859
Global name '_LE' was missing in trove/cmd/guest.py. This patch
adds this missing import.
Change-Id: Ia0b4ecb27f0f4dcae88a56288bc5052f1afa63e8
Closes-Bug: #1620888
Add --incremental flag to backup-create and set the default
value to False. When the flag is passed it will create a
backup based on the last good known full/incremental backup.
If there is no previous backup a full backup will be created.
Change-Id: I6fcd60f6068dfb0145fa64248a2d8b3fdfc1004d
Depends-On: I96498180890b8b1692b1bac3b19988532f6eb077
Partial-Bug: #1607429
My pycharm error, unresolved reference os , in this file,
my environment is py3.4
Closes-Bug #1618755
Change-Id: I1306d77418a16d9addad657a262250399601db53
Use an empty string rather None as marker for bisect.
On Python 3, a comparison between None and strings now raises a TypeError.
Partially implements: blueprint trove-python3
Change-Id: Iaefc20bd1cc980732e0c8a075e570958335c17e1
Several LOG messages were found to be incorrectly i18n'ed.
This change fixes them
Change-Id: I6e7e2e3f202392dd30417a9070745e70d4819f36
Closes-Bug: 1618858
Depends-On: I924cfd1ca93bef7f3dc79149519df10305390779
Recent changes to api need to be reflected in api examples. This has
come to light because client 2.5.0 is now released.
Change-Id: I924cfd1ca93bef7f3dc79149519df10305390779
Couchbase management user (root) is completely distinct
from end-user (bucket).
Buckets and management are essentially two different services.
Application connections are to buckets via SDK client using bucket creds.
Administrative connections are via management REST API with 'root' creds.
The helper ping() method used to verify end-user connections cannot
be used to assert management (root) connections.
This patch set implements root management connection assertion
via REST API. It makes more root tests pass.*
* Restore root tests still fail because Couchbase
does not report root on restore which is actually a bug (1535148).
I have a fix but it is included in a larger patch set which
I don't want to push at this stage
(maybe if we have leftover resources towards the end of the cycle).
Change-Id: I6b2fd63e5d2761ceadc5c13da7443450f27b92cc
This commit cleans up a couple of issues with Trove's tox.ini.
1. something that was reported to me by Matthew Treinish who
observed that the tox jobs are oddly bundled together.
This is important because infra's reporting is unable to generate some
useful metrics and tracking of trove's jobs because of the way they
are tied together.
2. with the cleanup of docs, apidocs is no longer a directory and can
be cleaned up, similarly api-ref is a thing and should be handled in
.gitignore.
3. that py27 used to use testr, switched it to using ostestr. I've
retained it for coverage.
4. the sqlite db needs to be deleted between runs because we otherwise
fail (on a developer configuration) because some tests intentionally
appear to leave stuff around.
5. I've set all ostestr jobs to run with --serial because there are
clearly cases where multiple tests can attempt to setup the DB at the
same time and get into trouble. Till we can fix those (if we feel it
is worthwhile) it is more reliable to run ostestr in --serial mode. It
only takes 2.5 minutes. This includes the cover jobs which runs with
--no-parallel.
An infra change will follow that will enable testing jobs.
Change-Id: I0a971c6ef6c1a4669490f2aa0e67c795fe86e190
Closes-Bug: 1618098
Closes-Bug: 1618149