Except for nova/api/, nova/tests/, nova/virt/
In preparation for automated checking being enabled (N306)
Change-Id: I5af70ea13b470ce58695ef0a5e0ff9da3ac1b135
Added notifications for volumes have been added with tests.
This includes create/delete/exists events for volumes.
blueprint nova-notifications
Change-Id: I21b74974fac22c3621ccf7564dc5c0d339f8751a
If the user allocates the wrong address display
a message about incorrect address rather than a traceback.
Fixes LP: #957707
Change-Id: I8ccd3056f755afb9f873a25fd494428e481ff3a2
Signed-off-by: Chuck Short <chuck.short@canonical.com>
Fixes bug #704737
With FlatManager, ensure that the network info's dhcp_server value is not set
and use that key to decide in the compute service whether DHCP firewall rules should be added.
Change-Id: I8183a6fa3881adea1a09f3f1a29442e6b7a919ce
The get_disks() method is not used anywhere yet, and is not tested properly,
but it does not work as reported in bug #998089.
This patch add unit test for it and fixes it so it's now possible to
retrieve the list of device name associated with a libvirt domain.
Change-Id: If09fb683364174fe02cbcbec4c6bbcda991287d6
Signed-off-by: Julien Danjou <julien.danjou@enovance.com>
The old dns_domains table was encoded in the 'latin1'
charset and has its primary key length set to 512.
This is too long to be a valid pkey in a proper 'utf8' encoded
table and seems to cause a slew of errors like:
1) Dumping a database with mysqldump and trying to import it fails
because this table is latin1 but fkeys to utf8 tables (projects).
2) Trying to alter the old dns_domains table fails with errors
like:
mysql> ALTER TABLE dns_domains DROP PRIMARY KEY;
ERROR 1025 (HY000): Error on rename of './nova/#sql-6cf_855'....
This patch creates a new migration that re-creates the dns_domains
table with a shorter 'domain' column which works with the utf8
charset.
Fixes LP Bug #993663.
Change-Id: I2060d39cbdabd20f410ebd501160a04c51641755
The recommended cluster size [1] for qcow2 is 64k, which is the default
when invoking 'qemu-img create ...'. Remove the cluster_size option
from qemu-img invocation.
[1] http://www.mail-archive.com/openstack@lists.launchpad.net/msg11371.html
V2: Fix test.
Change-Id: Iafc290720a820cb5f57ded10a2f584fdc5cd5b49
1. Added validation for image metadata.
2. Renamed the method 'check_img_metadata_quota_limit' to
'check_img_metadata_quota_limit' since it also validates the image metadata.
3. Added unit test cases.
Change-Id: I0e9dad97c03070363d14977897701e146870e41a
It was implemented (with a small bug), but never used. Also, lean on
the xenapi fake over using stubs in the migration tests to ensure we
test more code.
Change-Id: I96a7222a017302332bd6a3600fa645a74c502944
In implementing adding 'image_meta' to instance notifications,
this was committed:
https://review.openstack.org/#/c/7309
commit 53adfd289b
However, it's buggy because an instance could have been deleted after we
built from it... and bin/instance-usage-audit uses an admin context with
which we cannot query glance.
This stores image properties with instances in nova as system_metadata
and notifications will use that data.
Fixes bug 997833
Change-Id: I50575969b5cb28adaae9a713e749dc486772c417
instance system_metadata was not being marked as deleted when an
instance is marked as deleted.
Fixes bug 997849
Change-Id: I876b402ca33f4574bb426615f6c467ded6908fc2
Windows agent requires an argument of either 'agent' or 'xentools' to
the 'version' command. All we care about is 'agent', so add it. The
unix agent happily ignores the arg.
Fixes bug 997805
Change-Id: Ic369c8a2850173057da9d3175a02b5864d7a6514
Use cfg's new helper method to find the paste deploy config file. The
basic behavior is "look alongside the config file" with a fall back to
the standard default config paths.
Change-Id: If7a9ac1d456f372e3f5f6c78e237a7b099d1a30e
* Remove the instance relationship and instance_id FK
* Add instance_uuuid column to volumes table
* Passed unit tests and devstack tests
Change-Id: Id598f1f1d7915d1af6bf3dd75e5819dce08aaa0f
Removed snapshot_instance method on base virt driver. It is
essentially an unused duplicate of the snapshot method.
Change-Id: I082075d8d6797bd60072d375f3ce16704ff6351e
/proc/meminfo may show wrong values for the memory when using Xen, so
this fix computes the memory quering libivrt.
Change-Id: I188e2d34bcee13954653b93b9b816cf4530b8859
Added image metadata to compute.instance.exists notification.
This information is a convenience to assist billing/usage systems
by providing information about the image the instance was built
from.
Change-Id: I27e8c048ed419858ba2eb3e20c25ac454b0419b6
based on comments from https://review.openstack.org/#/c/7269
Removed duplicate declaration from test cases.
Change-Id: I090cd3ba0aa2ed3720223deea5e44cfe71682fe0
This restores the DB update for image_ref immediately on rebuild, but
passes the old image_ref to the manager for use with instance.exists
notification.
Fixes the fix for bug 997245
Change-Id: I8cf6a8759e8fe605bb023048a320b15eee3bcd80
Updates the XenServer vmops.py _shutdown function so that it
avoids calling get_info/_get_vm_opaque_ref with the instance name.
Since _shutdown already has the vm_ref there is no need to make
this call.
This fixes an issue where 'RESIZE' was broken when resizing to
the same host.
Fixes LP Bug #997313.
Change-Id: Ic5a1a0f92df9a2b809aef20a250843bf48402d75
Joe Gordon