Alex O'Rourke [Sat, 12 Mar 2016 00:05:00 +0000 (16:05 -0800)]
LeftHand: Add default SSH timeout and key values
Currently if ssh_conn_timeout or san_private_key is not specified
in the remote_array object, the client creation fails. This should not
be the case. Instead, we can add default values to allow the creation
to succeed.
In addition, this patch removes some redudent lines of code and
moves a call to a more logical place.
Ivy Zhang [Tue, 15 Mar 2016 09:25:36 +0000 (05:25 -0400)]
Use get_by_args instead of get_by_host_and_topic
We already changed to use get_by_args, but in init_host_with_rpc,
it still called get_by_host_and_topic instead of get_by_args.
Which would cause restart a disabled cinder-volume service failed
Ryan Liang [Tue, 15 Mar 2016 01:39:45 +0000 (01:39 +0000)]
Revert "VNX: Set timeout for naviseccli"
EMC CI test failed.
It is not the fault of adding timeout option to naviseccli.
I made a mistake by adding wrong option format in code.
The command line should be
self.command = (navisecclipath, '-timeout', '120', '-address')
instead of
self.command = (navisecclipath, '-timeout 120', '-address')
Later, I will send out a new review for the right one,
after this is reverted.
Ryan McNair [Mon, 14 Mar 2016 17:37:35 +0000 (17:37 +0000)]
Fix volume migration VolumeType exception
Since Iabf9c3fab56ffef50695ce45745f193273822b39 modified the default
loaded attributes, the obj_attr_is_set check in finish_volume_migration
is now True for VolumeType. This should not get copied to the new
volume (since it's lazy-loaded based on the volume_type_id), so
this patch adds volume_type to the list of keys to ignore. To get
test coverage, test_finish_volume_migration is updated to use the
default expected_attributes and dict comprehension is updated so
it succeeds without glance_metadata being present.
Tom Barron [Fri, 19 Feb 2016 23:19:14 +0000 (18:19 -0500)]
Fix retype failure when original has no volume type
cinder.objects.volume.Volume.finish_volume_migration()
should skip the volume_type_id field when swapping fields
between "source" and "dest" volumes. The "dest" volume
has already been created with appropriate volume_type_id
and the "source" may have not have a volume type.
This commit also changes a LOG.error() to a LOG.exception()
in where finish_volume_migration is called in order to
show the exception traceback in its calling context.
Alon Marx [Thu, 3 Mar 2016 16:15:38 +0000 (18:15 +0200)]
IBM XIV/DS8K: Implement Replication v2.1
This patch implements cinder Replication version 2.1 as proposed in
https://blueprints.launchpad.net/cinder/+spec/replication
The driver implements:
1. freeze_backend - notify the backend that it's frozen
2. thaw_backend - notify the backend that it's not frozen
3. host_failover - failover a backend to a secondary replication target
neochin [Sat, 12 Mar 2016 18:11:46 +0000 (02:11 +0800)]
Huawei: Implement v2.1 replication
This patch implements the v2.1 replication in the HuaweiDriver.
All function in replication v2 are removed. A new function,
failover_host, is added to support failover between backends.
When error occurred on primary backend, you can failover to
the secondary backend. But it's strongly recommended to recover
the primary backend.
After failover backend, the operations on the secondary backend
are limited. A simple example, if you create a new volume on
secondary, after failback, it can't be used. So, You should not
do that, such as create, delete, extend volumes, etc.
But you can detach volume if it's in in-use status, then do
attach to vm to make replication volumes R&W.
The configuration for v2.1 is almost the same as v2. Except that,
only one backend configuration is needed. A simple example as
follows:
Gorka Eguileor [Mon, 14 Mar 2016 11:54:58 +0000 (12:54 +0100)]
Remove circular import to fix config generation
Patch https://review.openstack.org/278354 created a circular import when
added an "import cinder.db" line to cinder/api/common.py since
cinder/db/api.py imports cinder.api.common back, which breaks genconfig.
WenjunWang1992 [Mon, 14 Mar 2016 09:10:16 +0000 (17:10 +0800)]
Wrong comment line in quotas.py
The commonts are not suitable in the code below:
1)in function "show"
origin:
:param id:target project id that needs to be updated
change to:
:param id:target project id that needs to be shown
2)in function "delete"
origin:
:param id:target project id that needs to be updated
change to:
:param id:target project id that needs to be deleted
liuke2 [Sat, 20 Feb 2016 09:02:50 +0000 (17:02 +0800)]
Huawei: Creating hypermetro failed in the remote pool
Create the hypermetro failed when use the remote pools on the
huawei driver, the log print get pool id fail. The 'pool_id'
is not exist in the LUN property dictionary, so the 'pool_id'
should be modified to 'PARENTID' to save the remote pool id.
Patrick East [Fri, 11 Mar 2016 23:28:27 +0000 (15:28 -0800)]
Fixup stats key for replication in init_host_with_pc
We were looking for just ‘replication’ in stats, but that key isn’t
used by anyone and would always be false. We instead want to be keying
off of just replication_enabled.
Vincent Hou [Fri, 11 Mar 2016 21:25:18 +0000 (16:25 -0500)]
Add the key 'replication' and set the correct 'replication_status'
This patch fixes the minor issues following the implementation
of replication v2.1 for storwize driver. It adds the key
'replication' to make sure of that the cinder-volume service
can have a successful restart and sets the 'replication_status
'back to enabled for failback.
Doug Hellmann [Fri, 11 Mar 2016 21:06:11 +0000 (16:06 -0500)]
register the config generator default hook with the right name
No config generator hooks should ever be registered with a name that
belongs to another project. In this case, using oslo.middleware.cors
means that *every other project* that loads the middleware gets this
application's defaults when the generator is run on a system with
everything installed (such as a dev box with devstack). Use the name
of the app instead, to ensure that the defaults are only set when this
app's sample config and documentation are being generated.
Patrick East [Fri, 11 Mar 2016 18:27:50 +0000 (10:27 -0800)]
Fix for Pure drivers not checking full client version
We were only looking at major and minor versions… we needed to be
looking at the whole thing. This change allows for checking against
all three version parts, and for even differing schemes as time goes on.
Patrick East [Fri, 11 Mar 2016 17:59:56 +0000 (09:59 -0800)]
Fixup for Pure drivers cheesecake replication setup
We had a remnant of the past replication design left behind in the setup
methods which was causing the arrays to be setup with bi-directional
replication. With the switch to cheesecake we do not need the secondary
arrays to be replicating to the primary. This change removes the step
which would configure that.
Ivan Kolodyazhny [Fri, 11 Mar 2016 16:14:06 +0000 (18:14 +0200)]
Remove empty directories and unused files from unit tests
1. Files from cinder/tests/unit/var are not used anywhere
2. cinder/tests/unit/wsgi/ and cinder/tests/unit/xenapi/ contain only
empty __init__.py files
Alex O'Rourke [Fri, 11 Mar 2016 00:09:01 +0000 (16:09 -0800)]
Cleanup 3PAR/LeftHand failover_host exceptions
The 3PAR and LeftHand drivers were raising the wrong exception,
VolumeBackendAPIException, instead of InvalidReplicationTarget,
which was putting the host in an unrecoverable state,
Dinesh Bhor [Tue, 9 Feb 2016 11:20:26 +0000 (03:20 -0800)]
Fix 500 error if 'offset' is out of range
If large value is passed as offset to snapshots, volumes,
backups, consistencygroups and qos_specs list api then it
throws 500 internal server error.
Moved existing validate_integer() method from
cinder.api.openstack.wsgi.Controller to cinder.utils so
that it can also be used for validating offset param for
integer value in _get_offset_param() method and return 400
error if value is out of range.
Alex O'Rourke [Thu, 10 Mar 2016 21:33:20 +0000 (13:33 -0800)]
Fix up failover_host exceptions to preserve states
UnableToFailOver and InvalidReplicationTarget do not save the state
of replication_status currently. This patch adds host.save() in
order to write the change to the db.
In addition, the manager should honor the current replication state
of the host when InvalidReplicationTarget is raised instead of forcing
it into 'enabled' state.
Helen Walsh [Sat, 13 Feb 2016 12:12:47 +0000 (12:12 +0000)]
EMC VMAX - SnapVX and other snapshot improvements
Using snapVX VMAX3 for creating snapshot and volume from
snapshot. Improvements in wait_for_sync to determine when
the max number of retries has been reached.
John Griffith [Wed, 9 Mar 2016 23:43:14 +0000 (16:43 -0700)]
Move replication_status update to init_with_rpc
We were using host_init to read replication_status from the
driver and update the service entry in the DB. It turns out
that on a fresh install this doesn't actually work, because
while we have multiple init methods for the backend, the
Service entry isn't actually created in a fresh deploy until
AFTER init_host. The result was that in some cases we were
trying to update a column on a non-existent Service in the
DataBase.
This patch moves the replication_status updates for the
service into the init_with_rpc method. That method was
just a noop stub in the parent manager class, so we just
implement it in cinder.volume.manager and do what we need
with the replication update info.
In order to allow a cloud operator to use the policy based
user access control for these operations, a context during
these operations should be elevated before db operations.
After applying this change, the cloud operator can manage
policy for volume type operations like this.
1. To permit volume type operations for specific user,
add "storage_type_admin" role.
2. Add "admin_or_storage_type_admin" rule to policy.json.
"admin_or_storage_type_admin":
"is_admin:True or role:storage_type_admin",
3. Modify rule for types_manage.
"volume_extension:types_manage":
"rule:admin_or_storage_type_admin",
Code Review / openstack-build / cinder-build.git / log