Some content in the admin docs was incomplete,
and irrelevant to administrators.
Also add notes on a couple of missing common
capabilities:
- mount_snapshot_support
- availability_zones
Partial-Bug: #1816486
Change-Id: Ie55fd5bf80ca2ebacdb652c701aa9c213f758963
Signed-off-by: Goutham Pacha Ravi <gouthampravi@gmail.com>
(cherry picked from commit 36a7ec422b)
Updates user guide documentation to provide metadata info
to an access rules. Also add two new operations for access
rule metadata: 'access-metadata set' and 'access-metadata unset'
Change-Id: Ib6008c9dd4284b8c3434ffce61fb74f1db6c3d35
Closes-Bug: #1782859
(cherry picked from commit e2e764d976)
This patch add some notes regarding the use of scheduler when
creating share from snapshot in another pool or backend.
Change-Id: I8d904cb3aab4d653178948284571f471bd59ac1b
Closes-Bug: #1878244
(cherry picked from commit 6b8f416e1b)
Extend docs on running manila API with a web server (Apache, NGINX)
Add samples for mod_wsgi
Partially-implements: blueprint wsgi-web-servers-support
Change-Id: Ic756466e65ffd8b5fc6b8e8fb3eef4b8535b46d6
(cherry picked from commit c1912b70f6)
Added some samples on uWSGI config and some docs
on how to run manila with uWSGI for contributors.
Partially-implements: blueprint wsgi-web-servers-support
Change-Id: I8e68619b00255197b3d3cf6796c60d1a3a04ff36
(cherry picked from commit 7a1414977f)
DellEMC manila drivers have bugs in returning capacity in pool_stat.
The current powermax/vnx manila return size in MB and unity manila
returns size in bytes.
Change-Id: I1a2c28548d02eee6632313a41a5c4d8ab6ad9532
Closes-bug: #1890372
Closes-bug: #1890375
Closes-bug: #1890376
(cherry picked from commit 53f756d065)
The existing "manila-grenade" job relies on
devstack-gate, a deprecated project. The Grenade
project now has a native zuulv3 style job that
we can inherit and run manila's upgrade tests.
Manila's grenade tests are only going to run
API tests, and hence this grenade job doesn't
have to enable nova, cinder, glance, neutron
and swift. However, bug #1887835 prevents
us from disabling nova at the moment,
and nova requires glance, placement and neutron
to be deployed, so we'll be enabling these
services too until that bug is addressed.
Change-Id: I1636c612ac2475f7a00c0888ef62daa6c516eef2
Signed-off-by: Goutham Pacha Ravi <gouthampravi@gmail.com>
(cherry picked from commit 660355eb86)
When a share driver does not provide the capability
to create share group snapshots, the share manager
iterates over the shares in the group and invokes the
driver method to snapshot the shares. However, the
create and delete routines in this fallback path
did not include some attributes necessary for the
operation in many drivers.
Change-Id: Id5a9467247df1d8f0ec6dee3fae842ba673c34ed
Closes-Bug: #1888905
Signed-off-by: Goutham Pacha Ravi <gouthampravi@gmail.com>
(cherry picked from commit a398743264)
Switch to using the LVM driver in the grenade
job that allows us to add a minor data
path test to prove that upgrading manila
has no impact on data path connectivity
to resources created by manila.
Change-Id: I8588e8f988d85dc64e19e7a44a25c3dd0b776892
Signed-off-by: Goutham Pacha Ravi <gouthampravi@gmail.com>
(cherry picked from commit 8cccb73f46)
During maintenance, administrators may decide
to unmount shares, snapshots or remove them
entirely prior to cleaning up LVM share resources
in manila. The driver should not fail on deletion
of missing resources.
Change-Id: Ieaf37ec10db9a8bdce6bb195b76335fea9b2b52f
Closes-Bug: #1888915
Signed-off-by: Goutham Pacha Ravi <gouthampravi@gmail.com>
(cherry picked from commit e57809fd74)
Compression is not needed post [1], explicit
compression renders them non-browsable.
This commit applies to legacy style jobs that
will soon be replaced with zuulv3 native jobs;
however, fixing it isn't terrible, and we
can use this fix for stable branches
where zuulv3 native jobs may not come for a
while.
[1] http://lists.openstack.org/pipermail/openstack-discuss/2020-January/011875.html
Co-Authored-By: yatinkarel <ykarel@redhat.com>
Change-Id: Icd4f9dc435a807f6f729e4eb2d4df4b282ec79ac
Signed-off-by: Goutham Pacha Ravi <gouthampravi@gmail.com>
(cherry picked from commit 770e57ff31)
ONTAP 9.5 adds vserver cifs domain preferred-dc configuration
validation. The parameter 'skip-config-validation' is mandatory
for 'cifs-domain-preferred-dc-add' now and is set to false,
i.e. an invalid config will raise early.
Change-Id: I5eb97ba1b29ad24b296e12064f81eb93533c6b99
Closes-Bug: 1887643
(cherry picked from commit 3057d0643e8a93c4ae857159d96e94541ad0be4d)
Because ``service`` commands are deprecated in favor of
``systemctl` commands, the check to see if the service
command succeeds in the NFS helper used by e.g. the
lvm and generic back ends fails. The string sought to
indicate an error is never present.
Update the helper to use ``systemctl`` commands instead
and error check for them correctly.
Closes-bug: #1887695
Change-Id: Ibb1c15f997fa60e1ba5bfb7437502d8aa454d35f
(cherry picked from commit 391ec7dffa)
- do *not* skip fsck when extending
- have extend resize the filesystem itself
rather than making a separate execution
of ``resize2fs``
Closes-bug: #1887694
Change-Id: I67ac549e790322c56d02db4424ec01a5449e6633
Policy checks are skipped when looking up a share-type
by name. This causes share creation attempts that specify
a valid share-type to pass the early API check on share type
even if the share type named is private and not shared with
the user's project. The share creation fails later, but after
the database record for the share is created. Although the
operation fails with an ERROR, the share is stuck in CREATING
state.
Fix this issue by checking the user's project in the database
API just as we do for share type lookups by uuid.
Closes-bug: #1885956
Change-Id: If5fe32c155fe0861b3ed86b862335e062796056b
(cherry picked from commit f877deed51)
NetApp ONTAP share delete operation can fail sometimes when is triggered
immediately after migration cancelation on a overloaded NetApp backend.
Migration cancelation invokes "abort_volume_move" which is an asynchronous
API. If share delete operation is requested immediately after call the
former API, it fails because the "abort_volume_move" is still in progress.
Now NetApp cDOT driver checks, for a period of time, if the
``volume-move-abort`` operation has ended before report migration
cancelation success.
This patch squash the following commit that improves the release note
for this fix:
[NetApp] Updating the release note for bugfix 1688620
(cherry picked from commit a0dd86a98788f7e2d1ca55be26c2a3dea4e36f57)
Change-Id: I76e11fef27c9723f019cfdfdc6ea86878db78776
Closes-Bug: #1688620
(cherry picked from commit 0ee4140823)
With python3.7, the eventlet is breaking the ssl.py, so the https
is not working. This patch fixes it by changing the request library
(urllib by requests), the new library can be built over the
pyopenssl.py instead of ssl.py.
Closes-Bug: #1878993
Change-Id: I9c0b1f332ead25634f3dc3aebfdc8b51dfbc4178
(cherry picked from commit 29622725e4)
When the NetApp backend starts, it needs to know whether the
`revert_to_snapshot` support exist. So, it was retrieving the
licenses and checking if the `SnapRestore` is included. Using a
scoped account, the backend cannot retrieve that information,
though. So, this patch solves it by sending a fake operation
`revert_to_snapshot` that must fail. Analyzing the given error, it
sets the backend support field.
Closes-Bug: #1882590
Change-Id: Ib71a6cec939288498e48736f129fbfdacaabe9da
(cherry picked from commit 6f58af1ae3)
After fixing the uwsgi installation from source that was broken[1][2],
Manila jobs started to fail since no uwsgi wasn't found in the
expected path. This fix now uses 'which' command to search for
uwsgi pathnames in the system.
[1] https://bugs.launchpad.net/devstack/+bug/1883468
[2] https://review.opendev.org/#/c/577955/
Related-Bug: #1883468
Closes-Bug: #1883715
Change-Id: I8d8b2fe07d86899c694cb73a81087d25311d30a5
(cherry picked from commit fbcabd2c03)
It is no need to enable ACE when creating CIFS share, so
remove this process.
Change-Id: Id9d248857c9a02eca486185d41b1fc22520d0919
Closes-bug: #1841035
(cherry picked from commit 47b90fa8fb)
Update documentation to fit into the new manila quota keys:
- `share_replicas`
- `replica_gigabytes`
Closes-Bug: #1881921
Change-Id: I91fa2222743da06557ceb52fcbae0f8928d0752c
(cherry picked from commit 7bc4886a03)
If the cephfs protocol is enabled, clients
need to access the ceph daemons.
We also don't need to enable access to NFS
ports when not using NFS.
Change-Id: I077d12785f94c940716f0e479d43dbb29ddc3c3c
Signed-off-by: Goutham Pacha Ravi <gouthampravi@gmail.com>
(cherry picked from commit 341b5777b2)
This patch adds documentation in the user guide module to make it
fit into the introduced share network with multiple subnets change.
Change-Id: I829f2e915ba2e20f15af639cccff99b6bba33a24
Partial-Bug: #1843953
(cherry picked from commit 9ee7482697)
This patch updates the admin guide to make it up to date with the
introduced changes for share networks in Train release.
Change-Id: I6ae48e2f174e0197fdeeaca5fd5a96a23c8bb182
Partial-Bug: #1843953
(cherry picked from commit fee462c2f0)
Not all drivers support share shrinking.
This patch adds a user message when shrinking fails due to operation
not being supported.
Change-Id: I5137cd972ec7737d70865038d7b82cfe13b9566c
Closes-Bug: #1802424
(cherry picked from commit 853d4fbceb)
Unmouting of the volume from service instance sometimes fails because
Windows clients tends to check the availability of shares even if not
mounted and the volume is in use then.
Fix finds out processes via smbstatus and kill them before unmounting.
Closes-Bug: 1703581
Change-Id: I2cbf959d7cb849d3ead1655b1a0932fd428a9e9f
Signed-off-by: Goutham Pacha Ravi <gouthampravi@gmail.com>
(cherry picked from commit 2c9fbc75059ac724cd845317473b7df4a8dcdc35)
The NetApp driver is fixed to avoid the deletion of cluster
ipspaces when deleting a share server. Now the share server
deletion skip ipspace removal if the ipspace is one of the
following: Cluster or Default.
Change-Id: I24cda6be69ef9258c6c637be48be32b93d835e0a
Closes-Bug: #1880747
(cherry picked from commit dbe34c08e2)
Neutron can filter ports by subnet-id by itself, so:
- Don't send all ports in Neutron on every share-server removal
- Don't send fixed-ips fields (which are nearly half of total data)
Closes-Bug: #1879754
Change-Id: If42c9a0a662a62bbe2a02b6baee6ae13eedd8c97
Signed-off-by: George Melikov <mail@gmelikov.ru>
(cherry picked from commit d5fc7d2802)
Monkey patch the original current_thread to use the up-to-date _active
global variable. This solution is based on that documented at:
https://github.com/eventlet/eventlet/issues/592
Change-Id: Ifc6420d927c0ce9e04ff3b3253e81a474591e9bb
Closes-Bug: #1863021
(cherry picked from commit 5e9f694a5a)
When running IPv6 tests, we disable the creation of
the public network in devstack, and handle it in the
devstack plugin. So, we'll need to configure it
in tempest in the plugin.
Change-Id: Icf5f26397e16d7fbfae77aa246ac5e35581f441e
Signed-off-by: Goutham Pacha Ravi <gouthampravi@gmail.com>
(cherry picked from commit 47c8debfeb)
We currently use a legacy devstack-gate script
to create the bgp speaker, and peer required
to communicate to and from the tenant private
IPv6 networks.
Move this logic to the devstack plugin so that
we can invoke it automatically with the existing
devstack variable MANILA_SETUP_IPV6.
Change-Id: Iea90e3f04ae05e5783c3163bbf2d2dabd128c7b5
(cherry picked from commit 2277c20c02)