Address miscellaneous issues with Markdown formatting in comments which
are consumed by shocco when generating the online documentation.
Change-Id: I953075cdbddbf1f119c6c7e35f039e2e54b79078
A few Markdown-oriented issues were causing Docutils errors to
leak into the end-user docs on http://devstack.org
Change-Id: I51fa9698afb1bfb48596478d83bd1fdcd84ac52e
When 'keystone token-get' fails, the caller can't detect the failure.
This cause troulbe shooting a bit complicated.
Change-Id: I3c58c5fd0e92a87e87546ea797904e08646a1097
Closes-Bug: #1238412
The actual regex checks a result in python format and
because of the change in the bug 1132961, Jekins failed.
I have update the regex to work with the old result
and the new result.
Change-Id: I393e1358f99be5f20d9ac8b3e214355a453ecfcb
Closes-Bug: 1239726
- Added changes to stackrc for the Trove Repos.
- Added support to devstack for "install", "configure",
"init", and "run"
implements blueprint:trove-devstack-integration
Change-Id: Ib3f6daad33e629f764a174b80762c808ce8588e2
* Add basic support for hypervisor plugins in lib/nova_plugins
* Add lib/nova_plugins/hypervisor-docker to use Docker as a Nova
hypervisor.
* Add tools/install_docker.sh to install the Docker daemon and
registry container, download base image and import
* Configure Nova to use docker plugin
* Add docker exercise and skip unsupported ones
Nova blueprint: new-hypervisor-docker
Change-Id: I9e7065b562dce2ce853def583ab1165886612227
The cause of some gating failure looks like
because of getting ip address on instance.
However current exercise didn't log the return value.
In this commit, we add get_instance_ip function with
error hanlding support, and apply it on the execise.
Change-Id: I8e17ba68093faafe58a98eb780a032368eea38aa
1) Use the newer cell_type configuration option instead of overriding
the compute_api_class.
2) A nova-cells service is started for both the API cell (region) and
the compute cell (child). The screen names were both 'n-cell' which
caused both services to log to the same log file. The screen names have
been changed to n-cell-region and n-cell-child.
3) Security groups are not supported with cells. Multiple exercises are
patched to use the 'default' security group instead of creating a new
one. Additionally, if one uses the 'default' security group, do not try
to delete it, as one can never delete the default.
4) Disable the floating_ips and aggregates exercises when n-cell is
enabled, as they are not supported by cells.
5) Related to #4, disable the floating_ips tests within euca.sh
6) Update the README.md. No services need to be disabled, and one only
needs to enable the n-cell service.
Change-Id: I9782d1e3cda3c9dd3daefa15c043f5b06473cb87
In devstack, the exercises/euca.sh test the "VOLUME". it should always
check the volume which was just created. Because there may be volumes
left by the last failed euca.sh, then the variable VOLUME
VOLUME=`euca-describe-volumes | cut -f2`
will be assigned a list of volumes, which is not expected.
Fixes bug #1191424
Change-Id: Ie004c9cfeaef4a8a313866d21d45c7b41f1b2ba0
* exercises/euca.sh: The match filtering InvalidInstanceID.NotFound
is cleaned up to eliminate numerous leaning toothpicks, and now also
actually works. The previous version was incorporating literal
square brackets into the pattern where I think a character list
match may have been intended instead, but was not actually necessary
anyway.
Change-Id: I765a20000e7b6ba6deadce8cad3bbc9559fcdd0d
There are Quantum plugins that do not require the use of an agent.
This patch relaxes the constraint in quantum-adv-test.sh by adding
the ability to customize the logic that verifies whether the test
can run or not, depending on the Quantum plugin of your choice.
Fixes bug #1177904
Change-Id: I6e2c17f43d9e1475b2eb175cceb5107a83f9aa74
Fixes test for detecting when a volume has been successfully
attached to an instance to not just rely on the state being
in-use, but also be marked as "attached". The attachment
state will be displayed when https://review.openstack.org/#/c/27098/
goes through. The attachment state is not currently displayed
by euca-describe-volumes because of the extraneous data returned
by the API as per bug #1074901.
In the meantime the "attaching" status of the volume itself, rather
than the attachment state, suffices for the test to work properly, but in the long
term this will disappear as it is not a valid ec2 API state and
volumes will move straight to the in-use state.
Fixes bug #1170548
Change-Id: Id38f37e1a0efc991f60da35145f809d98b5f41cd
Referencing %NUM% must be a mistake, because with it networks are not
deleted properly. See other commands within quantum-adv-test.sh as
an example.
Fixes bug #1164018
Change-Id: Ib222d82a1b8187167606b2543bb74c28b8465b5a
This patch changes the default network name/floating pool from nova
to pubic as the name public makes more sense than nova for a public
network.
Change-Id: I4153750c08d1510dbad3051681e8c142a79fe1a1
Adds Q_USE_SECGROUP flag for quantum security group
- Added has_quantum_plugin_security_group method for each plugin.
- Set NOVA_VIF_DRIVER to the hybrid VIF driver for plugins with
iptables based security group support.
- Specifying device_owner type on debug port in lib/quantum and
quantum-adv-test.sh. This change makes apply quantum security
group fro debug port
Change-Id: Ifd155798912247d85a9765ef73a2186b929237b4
- Set by default SWIFT_REPLICAS to 1 since this is the most common use
case when using it in devstack.
- If we have swift_replicas to 1 launch the object container account and
proxy servers in foreground in screen.
- Allow any s- services if we have just 'swift' in service enabled
instead of having to specify all s- services to enable swift. This will be
removed in the future.
- Set object server starting at 6013 instead of 6010 to not conflict
with remote ssh x11 forwarding.
Change-Id: I890b6953b70283bfa0927fff0cf5e92f3c08455b
This reverts commit 5a5cbf7274a6a50bb766ec590cf885430ed5c5d0.
This breaks on a default openstack install on Ubuntu 12.10,
revert for now until this can be cleaned up to work out of the
box.
Change-Id: I185509cc30bd28e920cdab60fc92129949bd6b0d
It is hard to grep errors in current log. so in this patch,
I'm updating die function which also writes log for
screen_log_dir/error.log.
In future, we may categolize negative fault by using
this error.log.
Change-Id: I70a8cfe67ed408284f5c88c762c6bb8acb8ecdb2
- Set by default SWIFT_REPLICAS to 1 since this is the most common use
case when using it in devstack.
- If we have swift_replicas to 1 launch the object container account and
proxy servers in foreground in screen.
- Allow any s- services if we have just 'swift' in service enabled
instead of having to specify all s- services to enable swift. This will be
removed in the future.
Change-Id: I496f79e14f99bd7e9f2c7deee12a4b6e935c3a5b
Fixes bug 1131244
aggregates.sh was making assumptions on the working directory. Thus it
failed, if the current working directory is not devstack. This pathch
makes the sourcing to use the full path for openrc.
Change-Id: I10c83c88112ead97b43873e1034b0e76e4f889f4
This should fix the failing (but non-voting) postgres gate tests.
Why does postgresql change the value '0.0.0.0/00' to '0.0.0.0/0'?
Clearly the correct value for the network CIDR bits is with only
one zero but even an incorrect value shouldn't be changing.
SQLalchemy is given this for the column:
Column('cidr', String(length=255)),
Change-Id: Ib19dad23789654664e90518087e5a462fa8b8034
A logic error in volume exercise's wait for volume deletion
causes the test to do the opopsite, and continue on even tho
the volume is in in the 'deleting' state. If using a volume
backend that can quickly delete volumes (ceph), and the volume
is gone before entering the wait, the loop will spin, timeout
and fail the test.
Change-Id: I8e3d2aaa04e6a165e0dee32bedac97d35e13d5eb
The behaviour of this case changed with bug/836978. Requesting the
status of an invalid instance will now return an error message
including the instance id, so we need to filter that out.
Resolves the devstack elements of bug 836978.
Change-Id: I385eb6f04cff90e1ddc0b79d835fbbdf92e4e9ff
this spuriously fails in devstack gate, and need to turn
this on to get to the bottom of the fails.
Change-Id: I7d5c1b4d3230efacdd8a3b89e5e40e98ac894a1d
Instead of implementing availability zones in the service table,
availability zones will be implemented using general aggregate metadata.
So when an aggregate is created it will already have metadata.
Part of blueprint aggregate-based-availability-zones
Change-Id: I0fd22399b99a14087fef63fc91d0baef746efbed
Fixes bug 1087656
In euca exercise, the timeout for one of the volume operations was
specified as ASSOCIATE_TIMEOUT, whereas the timeout error message was
mentioning RUNNING_TIMEOUT. This fix changes the timeout to
RUNNING_TIMEOUT so that it is consistent with the error message.
As RUNNING is usually larger than ASSOCIATE, it leaves more time for the
volume operation.
Change-Id: Ic016c7920ae6e4ec9a476bb5612b7df9eed01c75
sudo is only allowed in stack.sh on the CI, so move setup_quantum code
to the stack.sh.
also fixes quantum debug command setup for linuxbridge and ryu
Change-Id: I11bc0aa242a690e25acc088b3e9f483ceab38f26
from time to time horizon gets broken in devstack and it
takes a while to figure out why. Put a sanity check into
devstack exercises that checks for horizon front page being
up and not just a stack trace to use as a simple gate.
Change-Id: I13a6c59881f618d5194b1625b67115013c9cb6c2