| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| ... | |
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Now that the tgtadm patches are working, one problem was detected
The delete target operation would fail, and as a result lvremove
would fail due to an open connection.
There were a number of issue using the config file method, but
the most reliable way is to just use the iqn since we're going
to perform an os.unlink on the config file anyway.
devstack volumes.sh tests now pass and the detach/delete
calls are succesful.
Adjust test_iscsi for modifications to delete
Change-Id: Ic34ca2194b8d75fb84a06dfba6793106eb8055fe
|
| |\ \ \ \ \ \ \ \
| |_|/ / / / / /
|/| | / / / / /
| | |/ / / / /
| |/| | | | | |
|
| | | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Change the terminate_instance method of the compute
rpc API to take a full instance over rpc instead of just
the instance UUID. This cuts down on database access needed
by nova-compute.
This is the final change for instances in rpcapi.py. \o/
Unfortunately, I discovered that the scheduler isn't properly using
rpc API versioning when talking to compute nodes, so some of the methods
used by the scheduler still need some love.
Part of blueprint no-db-messaging.
Change-Id: Ic62450346ea90b9a6d64e282e992a21a4dd6ab64
|
| |\ \ \ \ \ \ \
| |/ / / / / /
|/| | | | | | |
|
| | | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
The existing code used trycmd(readlink) precisely because
/etc/selinux might not exist. The presented traceback in
the bug is only debug info.
Also the change will not work with libguestfs for example,
where the nova user can't access mount points mounted by root.
After discussion on IRC it was decided to reduce
debugging output in a subsequent commit.
Reverts: 9bea239f
Change-Id: Id07eae5b330332f36ad7caa786593e486aa36469
|
| |\ \ \ \ \ \ \ |
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
It turns out people have found ways to use the info cache everywhere.
The problem is when these ways require information about deleted
instances the info_cache may not be accurate.
This patch prevents the info_cache from being altered after it is
marked as deleted and marks an instance's info_cache as deleted
towards the beginning of the destroy process to ensure it doesn't
get blown away by a cache refresh somewhere along the way.
Change-Id: I0d5c2f1fc7e5f8f6c1154958986bb775ced07443
|
| |\ \ \ \ \ \ \ \ |
|
| | | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
Change 'r+' to 'r' for open() when just reading.
The '+' is uneeded and distracting.
Change-Id: I8c611ab6d63e165f86e588992011f73f69565249
|
| |\ \ \ \ \ \ \ \ \ |
|
| | | |_|_|/ / / / /
| |/| | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
Updates the ordering of the compute managers get_diagnostics
call so it works correctly with the current wrap_instance_fault and
checks_instance_lock decorators. To work correctly 'instance' needs
to be the first argument after the context.
Fixes LP Bug #1031788.
Change-Id: Ia7105eb272cd2932fc4f25b297ecab9e5ab8dbe2
|
| |\ \ \ \ \ \ \ \ \ |
|
| | | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
The core problem is that XenServer's `VDI.copy` call drops the
destination file directly into the SR. This means that half-completed
files are visible with no way to distinguish these from fully-copied
files.
We had some code that attempted to mitigate this issue by checking
physical_utilisation against an expected value. The problem with this
code is that it didn't account for VDI chaining where the
physical_utilisation would not necessarily match the parent.
The net effect of this was that 'cloned' VDIs would never be found
because their physical_utilisation was far below what was expected.
The work around is to create our own `_safe_copy_vdi` which is isolated
and atomic. Long term, `VDI.copy` should be fixed so that half-completed
files are never stored in the SR.
Change-Id: I6eb3cb5259f9ee1c7394e58f76105a8b39bfc720
|
| |\ \ \ \ \ \ \ \ \ \
| |_|_|/ / / / / / /
|/| | | | | | | | | |
|
| | | |_|_|_|_|_|/ /
| |/| | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
Add check_exit_code to check the ip address
creation in ensure_vlan.
Ignores error code '2' and '0' and '254'
from iproute2, which means the address
has been added or removed.
A further audit of return codes will come later.
LP: #1011640
Change-Id: Ie95b106446c9f136fa079d95df7a14fe50046696
Signed-off-by: Chuck Short <chuck.short@canonical.com>
|
| |\ \ \ \ \ \ \ \ \
| |_|_|/ / / / / /
|/| | | | | | | | |
|
| | | |_|_|_|/ / /
| |/| | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Fixes LP 1022961
Previously an attempt to set the root password would fail silently if
for example the driver did not support this action.
To avoid the user being misled that the password change had succeeded,
we now report any failure by changing the MEP for set_admin_password
RPC from cast to call.
Change-Id: I1fd7b925e9226d30892f3a7cc4ddb938ff678a55
|
| | |_|_|_|_|/ /
|/| | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Updates the network manager to use db.instance_get instead of
db.instance_by (which doesn't exist). This change seems to have
slipped in in the recent vifs UUID conversion code.
Fixes LP Bug #1031756.
Change-Id: Ie487a2f845a722a5e7923e26a550d74f553e8afc
|
| |\ \ \ \ \ \ \ |
|
| | | |_|_|_|_|/
| |/| | | | |
| | | | | | |
| | | | | | | |
Change-Id: Id7c4b0c4f8710652249b5c4fcb82abd5cccde6dd
|
| |\ \ \ \ \ \ \
| |_|_|_|/ / /
|/| | | | | | |
|
| | | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Updates the nova api-paste.ini file with a custom signing_dir
for Nova. This works around potential multi-service permission issues which
can occur when multiple OpenStack services share a common signing directory.
Change-Id: I99df63838fd41a4cade99b54da9511978b0d1a86
|
| |\ \ \ \ \ \ \ |
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
* Add comment explaining relevance of _is_image_available
* Drop superfluous log statements
* Drop GlanceImageService.delete_all as it has never been implemented
* Drop the 'features' kwarg for GlanceImageService.update
* Code simplification
Change-Id: Ie538dd93dec729fb060bbcddfbddcba6cf303e52
|
| |\ \ \ \ \ \ \ \
| |_|_|_|_|_|/ /
|/| | | | | | | |
|
| | | |_|_|/ / /
| |/| | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
tgt-admin needs to use --execute --update to tell
tgtd to re-read its configuration files when target
is created.
(LP: #1031509)
Change-Id: I71fe4d90a8cc9bcf7a4817aefd6de52e216102a4
Signed-off-by: Chuck Short <chuck.short@canonical.com>
|
| |\ \ \ \ \ \ \ |
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Change the change_instance_metadata method of the compute
rpc API to take a full instance over rpc instead of just
the instance UUID. This cuts down on database access needed
by nova-compute.
These methods were moved in rpcapi.py and test_rpcapi.py but only because
the rest of the file was in alphabetical order. The move puts it in the
right place to maintain that order.
Part of blueprint no-db-messaging.
Change-Id: I7dd2fb7480010893f66c753b3d31a4556009efe0
|
| |\| | | | | | | |
|
| | |/ / / / / /
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Change the unrescue_instance method of the compute
rpc API to take a full instance over rpc instead of just
the instance UUID. This cuts down on database access needed
by nova-compute.
Part of blueprint no-db-messaging.
Change-Id: I9a693b347ce583aa350f4c10531b53e0f9813577
|
| |/ / / / / /
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Previously the decorator current called out indirectly to
nova.api.network.api.API.get_instance_nw_info(), which is also
decoratored by @refresh_cache.
Change-Id: I8b71e2f6c85d007cde0da32890168c922102a7d8
|
| | |_|_|/ /
|/| | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
If a rebuild or a resize fails, then we should error out the instance.
This code takes an idiom we already use in the compute manager and adds
it to the rebuild and resize code.
Change-Id: Ia90c038eb4677e109d58e9b79bb8a05490689c9f
|
| |\ \ \ \ \ |
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
We can just use members of the instance dictionary without turning
them into local variables, as we don't need to populate instance_uuid
manually, as that is done by the @checks_instance_lock decorator.
Change-Id: I2046d5688c2f5122624dc338680b7ddd850ed330
|
| |\ \ \ \ \ \ |
|
| | | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Fixes bug 1031405
The listeners kwarg is deprecated in SQLAlchemy 0.7.0. Since 0.7.3 is
the minimum required version of SQLAlchemy, use the new event.listen()
API instead.
Change-Id: I3d2bd2e65dbc6787d1c9b69ffd0bcff85c157869
|
| |\ \ \ \ \ \ \ |
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Now that we're passed through an instance, or something which is
immediately turned into an instance, we should prefer using the full
instance inside compute methods. This is especially useful for log
messages, as it allows users to format the messages more than they
can if only the UUID is passed.
Change-Id: I29370f33a63817f3e45d453499bd46131b947d7e
|
| |\ \ \ \ \ \ \ \ |
|
| | | |/ / / / / /
| |/| | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
Addresses LP 1029911.
Associating an already-associated floating IP with another instance
results in the floating IP being first disassociated from the original
instance and then associated with the second instance.
However when this transfer of the floating IP to the second instance
occurs, the network info held in the database instance info cache was
previously only updated for the second instance (via the
@nova.network.api.refresh_cache decorator).
As a result, the floating IP was initially reported as being associated
with *both* instances. This misreporting is temporary - eventually the
periodic heal_instance_info_cache task will roll around and purge the
inaccurante network info for the original instance. However depending
on the frequency configured for this task and the size of the instance
population, it could take a while for the refresh to occur.
This issue is addressed by immediately purging the obsolute network info
when the reassignment occurs.
Change-Id: Ibfb04997ff4a983d144d81350857cb07475aff54
|
| |\ \ \ \ \ \ \ \
| |/ / / / / / /
|/| | | | | | | |
|
| | | |_|_|_|_|/
| |/| | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
The VHD sequence validation code was erroneously counting `swap.vhd`,
which caused it to raise an exception when a corresponding numbered VHD
was not found.
The fix is to simply ignore the `swap.vhd` file.
Other unknown VHDs will generate an exception, but from a
sanity-checking perspective, this is a Good Thing(tm).
Fixes bug 1030939
Change-Id: Ic82ae27a4af7ea8f7669fd006aea1a310b691218
|
| |\ \ \ \ \ \ \ |
|
| | | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
don't throw exceptions if %(color)s tag is used
fixes bug #1030078
If the log format strings include the %(color)s key in them and you
enable file logging, your log files won't work, and instead you'll
see exceptions on the console as the default python log handler doesn't
support that keyword.
This approach stubs out the extra keywords to be empty strings to avoid
the KeyError. Tests are also added to demonstrate the KeyError as defined
behavior, and test that color codes can be passed.
Change-Id: I24b2f45dd5a38b27b8cc5a70e0084d367e218172
|
| |\ \ \ \ \ \ \ \ |
|
| | | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
In the case where the command fails, we're returning
the err anyway and the caller can log if required.
In the discard_warnings=True case, then the caller
has explicitly asked to suppress the warnings,
so don't log this noise either.
Fixes bug: 1030609
Change-Id: I075c21aa6508182ae6dd030cf41ec31916559957
|
| |\ \ \ \ \ \ \ \ \
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
* changes:
Fix arg to get_instance_volume_block_device_info().
Send a full instance in snapshot_instance.
Send a full instance in set_admin_password.
Send a full instance in rollback_live_migration_at_destination.
|
| | | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
Michael Still caught this bug while reviewing an earlier patch in this
patch series. This code was passing an instance ID to this function,
but it expects an instance UUID.
Not really related to, but tied up in the patch series
for blueprint no-db-messaging.
Change-Id: I3ec871deadd13a8117f16ad46c31b0a9c8a016db
|
| | | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
Change the snapshot_instance method of the compute
rpc API to take a full instance over rpc instead of just
the instance UUID. This cuts down on database access needed
by nova-compute.
Part of blueprint no-db-messaging.
Change-Id: I401beb932e329683c3434ecd643db345aa217492
|
| | | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | |
| | | | | | | | | | |
Change the set_admin_password method of the compute
rpc API to take a full instance over rpc instead of just
the instance UUID. This cuts down on database access needed
by nova-compute.
Part of blueprint no-db-messaging.
Change-Id: Iddcc7cb068090faa98f0bb87307e09d5b0ebf0c2
|
