| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit 2c63cae921d024f39f7d14459873d20caa5cf9e0 added some havana
specific rpc handling to aid with the live upgrade from havana to
icehouse. Unfortunately, the processing was broken for some
methods. It would claim to be sending a '3.0' message, but would
be sending some of the arguments in an old format, causing some
methods to break when processed on the remote end.
This patch fixes up the logic to ensure that if messages are sent as
'3.0', that the format of the data sent in the message matches.
Note that this is a stable/havana specific change.
Change-Id: I54135effe4e083f73cb7d912ac9e82fd182e2b2d
Closes-bug: #1301436
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In an HA deployment, a 60 seconds delay between reconnects can be quite
problematic. This patch changes the delay calculation by setting the max
delay to 5s and by changing the way it is increased.
Unfortunately, this is one of the places where both our main drivers are
not consistent. Rabbit's driver uses configuration parameters for this
whereas qpid's driver has never had one. However, I would prefer not
adding configuration paremeters to qpid's driver for the following
reasons:
1. Most of OpenStack services depend on the messaging layer, hence
they need it to be available. A 5s delay seems to be reasonable and
I could argue the need of tune it further. Although so frequent
reconnects can add load to the network, that wouldn't be the main
issue if one of the brokers go down.
2. We're trying to move away from configuration options towards using
transport URL. This path is still not clear and I would
prefer avoiding adding new options until we clear it out.
Closes-bug: #1281148
Change-Id: I537015f452eb770acba41fdedfe221628f52a920
(cherry picked from commit 8b628d1e024f787dbb93d508117d9148388c0590)
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This is a backport of 4 fixes squashed into one because:
1. They need to all be merged together as they build on each
other as problems were found in each change after it was
merged on master.
2. The 3rd change won't pass Jenkins on it's own so it has
to be squashed with the 4th and final change, so I'm just
going to squash the entire topic branch together.
The change bugs fixed and cherry pick commit hashes are left
intact for each change.
Closes-Bug: #1241275
(cherry picked from commit 51e5f52e4cb60e266ccde71f205c91eb8c97b48b)
Cache Neutron Client for Admin Scenarios
Closes-Bug: #1250580
(cherry picked from commit 85332012dede96fa6729026c2a90594ea0502ac5)
Users with admin role in Nova should not re-auth with Neutron
Closes-Bug: 1250763
(cherry picked from commit 1c1371c78b990447aeaa4377b512f8887e6ff3ce)
Fix Neutron Authentication for Metadata Service
Closes-Bug: 1255577
(cherry picked from commit 652620d12f3afe6845e41d9762b52d23f44fd557)
============
Change-Id: I2858562b180f3e058a2da9d67bef02af80927177
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Use pkg_resources to determine correct boto interface
to use in API tests. stable/havana version.
Fixes Bug #1239220
Rebased-By: Dirk Mueller <dirk@dmllr.de>
Change-Id: I455ce5f2aabfad7178be361844c6b9b7b0d9c946
|
|\ \ \
| | | |
| | | |
| | | | |
stable/havana
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
post_live_migration_at_destination make libvirt.xml to destination
host. when execute (live) block migration for vm made by bootable
volume, method post_live_migration_at_destination write a
libvirt.xml in destination host.
but it missed block disk information so moved libvirt.xml always
has a wrong disk information.
I add block_device_info in blockinfo.get_disk_info at
post_live_migration_at_destination.
This change will make libvirt.xml with valid block disk information.
Closes-Bug: 1271780
(cherry picked from commit 0486647e1c85947bc48cbdd279b3f49824a6e692)
Conflicts:
nova/tests/virt/libvirt/test_libvirt.py
nova/virt/libvirt/driver.py
Change-Id: I31b3f2ef5f03f9ef61524bdb7ae089b18e4cd254
|
| |_|/
|/| |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The attack is based on creating a raw image that looks like a qcow2
image, and taking advantage of the code that used 'qemu-img info' to
autodetect the image format.
Now we store the image format to a 'disk.info' file, for Qcow2 and Raw
images, and only autodetect for images that have never been written to
that file.
SecurityImpact
Conflicts:
nova/virt/libvirt/imagebackend.py
Manual tweaks to some mocking in test_imagebackend.py
Change-Id: I2016efdb3f49a44ec4d677ac596eacc97871f30a
Co-authored-by: Nikola Dipanov <ndipanov@redhat.com>
Closes-bug: #1221190
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
From time to time an exception is raised in this method causing
the nw_info cache not to be saved. If this occurs we should raise
as this error will cause later errors to occur. For example, one
won't be able to associate a floatingip with the instance as there
is no nw_info found in this table. In addition, the fixed_ips on
the instance won't be returned via the api.
This patch also stubs out update_instance_cache_with_nw_info in a
few tests where an exception was being raised previously but went
unnoticed as it was not reraised but now is.
Related-Bug: #1252849
Related-Bug: #1249065
Change-Id: Ic860f72210ba736e11c10df21c4cb7625e9c0928
(cherry picked from commit 4c03383f21bc13caf3fed4db5aa1317d37758d5c)
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The nw_info value [] was being incorrectly equated with None. This caused
infinite recursion when api._get_instance_nw_info returned [] and was wrapped
with @refresh_cache (e.g., _get_instance_nw_info in
nova.network.neutronv2.api.API).
Fixes bug #1251792.
Change-Id: I3034fbb20e790f2d6f22c50b74a9f9dcdf9081aa
(cherry picked from commit 8bc974be50bea6ff6beacda890d9a33fa90059b3)
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The hypervisor version is expected to be an integer and not
a string. The nova/utils.py change is part of the following
patch but is not being backported as it adds new feature
functionality and is therefore not acceptable for backport
https://review.openstack.org/#/c/36073/
Closes-Bug: #1195139
(cherry picked from commit e2efa57ab2fb03b92f80abbf612a257c28475140)
Conflicts:
nova/tests/virt/vmwareapi/test_vmwareapi.py
nova/virt/vmwareapi/host.py
Change-Id: Ifac9ca3480191413e3a02f0dae94e452707ae82f
|
|/ / /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Migration script 216 will wipe out all data in quota_usages table
when user_id is null. After the upgrade from Grizzly to Havana is
completed, the migration script 203 adds a new column user_id, so
user_id is null for all data in quota_usages. So the correct
behaviour should be sync quota_usages table first and then delete
data if user_id is still null.
Fixing the existing migrate script is because we can not find
the quota_usages data after migration script 216.
Closes-bug:1245746
Change-Id: I4db62884785a4eb6e1db2c5b114b9c6e1b4882ad
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
If a snapshot is attempted to be taken when Glance can't find the
underlying image (usually because it's been deleted), then ImageNotFound
is raised to nova.compute.manager.ComputeManager._snapshot_instance. The
error path for this leaves the instance in a snapshotting state; after
which "operator intervention" is required.
This simply reverts the instance state under that specific circumstance.
DocImpact: previously a snapshot that failed for this reason would leave
the instance in a spurious snapshotting task state indefinitely. Now,
the task state is reverted. At the moment that revert is 'silent'
(although the failure to snapshot will be evident to the user).
Fixes-bug: #1267097
(cherry picked from commit 96dba50737819c24b4ddd9cb1078da305638abc9)
Conflicts:
nova/compute/manager.py
Change-Id: I9623cfaed31e8337ab65bc1eab85ad3b51b1a5ca
|
|\ \ \ \ |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Previously, in the setup method of the libguestfs driver
for Nova's VFS, on an error the handle object was simply
dereferenced (set to None), and then an new error was thrown.
This relied on an implicit close() on the handle being called
by GC. However, in some cases the setup progresses far enough
that the implicit close is not enough, and leaves an "orphaned"
libguestfs VM.
Now, the teardown() method is called explicitly, which ensures that
no "orphaned" VMs are left around in case of an error.
Change-Id: I7fbe470046ce6c76bc13d77d8df360351a3ef715
Fixes: bug #1270304
(cherry picked from commit fa62cfe15fd6dddecf657966d166a2d3a2bdd317)
|
|\ \ \ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
There is small chance for VNC port collisions with the current
implementation which choose the port number based on the MoRef id
of the VM.
This patch fixes this by running a query for all allocated ports
and then selects one which is not taken.
Closes-Bug: #1255609
(cherry picked from commit 2f49ed4b5dbb5c954fc7a9b42ee7b170c38c775c)
Conflicts:
nova/exception.py
nova/tests/virt/vmwareapi/test_vmwareapi.py
nova/tests/virt/vmwareapi/test_vmwareapi_vm_util.py
nova/virt/vmwareapi/vm_util.py
nova/virt/vmwareapi/vmops.py
Change-Id: If7c3b14dd49ed05c5fde819c5a36d5608650cbbc
|
|\ \ \ \ \ \
| | | | | | |
| | | | | | |
| | | | | | | |
file" into stable/havana
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
This patch sets secret=True for the following config option
neutron_metadata_proxy_shared_secret
Conflicts:
nova/api/metadata/handler.py
Change-Id: I381ea9e89f76beac11b03def6e3ccfee2a55e48f
Closes-bug: #1288463
(cherry picked from commit fa73a5d037471dbec11893ad83bd93276968b9af)
|
|\ \ \ \ \ \ \
| |_|/ / / / /
|/| | | | | | |
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Rbd.libvirt_info function definition misses hypervisor_version argument added in change:
https://review.openstack.org/32379
Closes-Bug: #1233188
Change-Id: Ib68d743e783af0f6d82d2ba180869ee642e86050
(cherry picked from commit 7a34be0ec0cd0cb9555fe64ff6c486faae1ae91d)
|
|\ \ \ \ \ \ \ |
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
A snapshot of a instance with an attached volume will result in
the wrong volume being uploaded to glance. The reason for this
are as follows:
1. the last image in the list of images would be exported
2. the base image would not be exported but the first found
The same happens with a VM rescue if there is a cinder volume
attached.
The solution is to search for an image that contains the instance
UUID.
Change-Id: Ic85b0ffd4f1c34f10d07b1a198eaad2030803d6f
Closes-bug: #1243193
(cherry picked from commit 9ceee09594cd35c264cab3e3b3d1d80aaa3fdbc4)
Conflicts:
nova/virt/vmwareapi/vm_util.py
nova/virt/vmwareapi/vmops.py
Change-Id: I6b9febbfe4e4e6ecf5d31a2b09a4319b07696f9c
|
|\ \ \ \ \ \ \ \ |
|
| |/ / / / / / /
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Currently the cleanup_running_deleted_instances peroidic task
failed with InstanceNotFound exception, this is because the
db query is not including the deleted instance while
_delete_instance_files() in libvirt driver.
Closes-bug: #1280140
Change-Id: Ie65ff255ad3c582a71db93b07304a21d4976a193
(cherry picked from d835163759527a651f3c4f2109ca0fdc3e968d37)
|
|\ \ \ \ \ \ \ \ |
|
| |/ / / / / / /
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Multipathing for attach volume and detach volume operations
doesn't work properly if there are different targets
associated with different portals for a mulitpath device.
Closes-Bug: #1266051
(cherry picked from commit 429ac4dedd617f8c1f7c88dd8ece6b7d2f2accd0)
Conflicts:
nova/tests/virt/libvirt/test_libvirt_volume.py
nova/virt/libvirt/volume.py
Change-Id: Iab316526cf64de60325d90043f73fa3e83fafb4e
|
|\ \ \ \ \ \ \ \
| |_|_|/ / / / /
|/| | | | | | | |
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
If an instance is in the PAUSED state before the compute node is
restarted, it will be left in an inconsistent state (impossible to
delete and/or restart it) after the node boots again.
This patch adds a check in the periodic task sync_power_state so when
this situation is detected the instance is stopped and left in a state
from where it can be started again or deleted.
Closes-Bug: #1265494
Change-Id: I7266ace596598522c28a3839091ef50faf4463c2
(cherry picked from commit 6abda5b738c5d801ede3328cf0ec1dc9dddc022f)
|
|\ \ \ \ \ \ \ \
| |/ / / / / / /
|/| | | | | | | |
|
| |/ / / / / /
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
The exception handling in the Vim module was too broad and did not treat
the suds.MethodNotFound exception. This would result in a exception
that "__deepcopy__" is not supported in the VIM module.
In the case of an exception the VMwareAPISession class may be deleted.
The patch also ensures that the deletion is able to logout if necessary.
Closes-bug: #1228847
(cherry picked from commit be5d39eacc84c11f96357a624ffa1cdeda8a568d)
Conflicts:
nova/tests/virt/vmwareapi/test_vmwareapi.py
Change-Id: Ic85be9b3407c444db4b3a074108c01d9141c61be
|
|\ \ \ \ \ \ \ |
|
| | |_|_|/ / /
| |/| | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
When running the CI there were cases when the aformentioned method would
throw a DBDeadlock exception.
Change-Id: I98d1a804c51e1bf3bb96193d82ffe7e5d064e134
Closes-bug: #1253455
|
| |/ / / / /
|/| | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
If the _call_method fails, for example a resource is not found,
then the same call will be invoked again due to the fact that the
polling task was not stopped.
Closes-Bug: #1246848
Change-Id: If4818e41aed5be03395ab22afc0edf9abb3c34e1
(cherry picked from commit 1c1570ff7ac8470d7dba28c893300835b22e0966)
|
|\ \ \ \ \ \
| | | | | | |
| | | | | | |
| | | | | | | |
stable/havana
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
The following patch removes an unneeded call to conductor to retrieve
data already available in instance.
Note: six.text_type() was added and wrapped around info_cache in some of
the tests. This was done because values from the db are in unicode
(text_type). The reason the test changed is because previously we
did not check the type of string from what the conductor was returning.
Closes-bug: #1258360
Conflicts:
nova/tests/network/test_neutronv2.py
Change-Id: I45a24af11534911be68fd0964b79fdf726d88de6
(cherry picked from commit 10004672ad1476c55deaad53684a50358da6f656)
|
|\ \ \ \ \ \ \
| |/ / / / / / |
|
| | |/ / / /
| |/| | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
get_instance_nw_info calls _get_instance_nw_info,
which is decorated with @refresh_cache. This in
turn calls update_instance_info_cache, again. This
is both expensive and more importantly dangerous because
the method update_instance_info_cache may call
_get_instance_nw_info itself, thus causing an infinite
loop.
Change-Id: Id8f0241d1c6609cf4e064ee6855228ae7de250e2
Partial-bug: 1235435
(cherry picked from commit 651fac3d5d250d42e640c3ac113084bf0d2fa3b4)
|
|\ \ \ \ \ \
| |_|_|/ / /
|/| | | | | |
|
| |/ / / /
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Reservations can deadlock when concurrent commits are run from
multiple hosts on the same project. This fixes the issue by using
our convienient retry on deadlock feature.
Change-Id: I31d74f4267eaa33e798de926069d1a0f63006a59
Closes-bug: #1274341
(cherry picked from commit 9c556d1557a7056d6056b25d3b91ff4abcb97971)
|
|\ \ \ \ \ |
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Although this code error is part of the `ping_listener` tests, it
missing from the list of known - and ignored - connection errors.
The code error 2013 refers to a connection lost during a query, see:
https://dev.mysql.com/doc/refman/5.0/en/error-messages-client.html#error_cr_server_lost
Without this error code, the session won't try to reconnect to mysql,
which makes `max_retries` useless.
Closes-bug: #1276510
Change-Id: If2667cf83ac656666ffe00aa90e3454516b080db
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
In the event that thete is no space on the datastore then ensure
that the disk extension is cleaned up accordingly.
A number of files would be left behind on the datastore.
Change-Id: Ica9daf9151e4fb1874fa8b1d7247080860ef4ae2
Related-Bug: #1255355
(cherry picked from commit 787d36b40b41c1ee1e56ae3f8d17984e70c59a8e)
|
|\ \ \ \ \ \
| |/ / / / /
| | | | | |
| | | | | | |
stable/havana
|
| |/ / / /
| | | | |
| | | | |
| | | | |
| | | | | |
Closes-Bug: #1244092
Change-Id: I17882e407afa00167e550a565081219f9e432a31
|
|\ \ \ \ \ |
|
| |/ / / /
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
If an exception is raised while get a libvirt domain's vcpu info,
the update_available_resource periodic task will be failed, which
will result in the resource of this compute node will never be
reported.
This patch add an exception handling to avoid this situation.
Closes-bug: #1270008
Change-Id: I69109402416989fdaa421f8dbc72953bd067c407
(cherry picked from commit 844df860c38ac38550b8d1739fd53131cd7fd864)
|