Bug 1161413

Summary: [RFE][cinder]: NFS driver snapshot support
Product: Red Hat OpenStack Reporter: RHOS Integration <rhos-integ>
Component: openstack-cinderAssignee: Eric Harney <eharney>
Status: CLOSED ERRATA QA Contact: Tzach Shefi <tshefi>
Severity: low Docs Contact: Don Domingo <ddomingo>
Priority: medium    
Version: unspecifiedCC: aguetta, cschwede, ebarrera, egafford, eharney, jomurphy, markmc, nlevinki, pgrist, rhos-flags, rkharwar, sclewis, scohen, sgotliv, slinaber, srevivo, tshefi
Target Milestone: Upstream M3Keywords: FutureFeature, TechPreview, Triaged
Target Release: 11.0 (Ocata)   
Hardware: Unspecified   
OS: Unspecified   
URL: https://blueprints.launchpad.net/cinder/+spec/nfs-snapshots
Whiteboard: upstream_milestone_none upstream_definition_approved upstream_status_needs-code-review
Fixed In Version: openstack-cinder-10.0.0-0.20170131044149.25a3765.el7ost Doc Type: Enhancement
Doc Text:
The NFS back end driver for the Block Storage service now supports snapshots.
Story Points: ---
Clone Of:
: 1515672 (view as bug list) Environment:
Last Closed: 2017-05-17 19:22:10 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---
Bug Depends On:    
Bug Blocks: 1515672    
Description Flags
Packstack answer file ocata.conf, cinder.conf, cinder logs none

Description RHOS Integration 2014-11-07 05:06:58 UTC
Cloned from launchpad blueprint https://blueprints.launchpad.net/cinder/+spec/nfs-snapshots.


Add support for snapshots to the NFS volume driver

Specification URL (additional information):


Comment 1 Sean Cohen 2014-11-26 03:33:03 UTC
The feature is under review for Kilo milestone 2 

Comment 2 Sean Cohen 2014-11-26 03:35:03 UTC
*** Bug 1155504 has been marked as a duplicate of this bug. ***

Comment 3 Sean Cohen 2015-08-24 20:06:53 UTC
Full Spec available at: http://specs.openstack.org/openstack/cinder-specs/specs/liberty/nfs-snapshots.html

Comment 4 Sergey Gotliv 2015-08-27 13:52:04 UTC
Eric, is leading this effort upstream. The spec file is already merged, but it probably will be implemented only in the M cycle.

Comment 5 Eric Harney 2015-08-27 14:26:41 UTC
This is currently blocking on a required fix in Nova:

Comment 10 Christian Schwede (cschwede) 2017-01-26 15:03:54 UTC
Upstream patches merged, moving from ON_DEV to POST.

Adding a needinfo - need to document some config change and limitations for anyone to actually be able to use it there too

Comment 11 Eric Harney 2017-01-26 16:17:51 UTC
Deployments must set "nfs_snapshot_support = True" in the volume driver backend section of cinder.conf for this feature to be enabled.

Comment 13 Tzach Shefi 2017-03-07 10:30:39 UTC
Eric, need you insight on this. 
There is a temp issue with some undercloud bug can't install with Infra, started testing this with Packstack. 

Simple AIO server 

On packstack answer file section:

On cinder.conf also enabled
# grep nfs_snapshot /etc/cinder/cinder.conf 
nfs_snapshot_support = True

Cinder create works fine, vol created on nfs backend
# cinder list
| ID                                   | Status    | Name | Size | Volume Type | Bootable | Attached to |
| e84c299d-fafd-4c4f-9a23-5ec3f1337057 | available | -    | 1    | nfs         | false  

Cinder snapshot create fails

[root@panther13 ~(keystone_admin)]# cinder snapshot-list
| ID                                   | Volume ID                            | Status | Name      | Size |
| e1c82a7c-3729-4098-998f-aeb620b27bce | e84c299d-fafd-4c4f-9a23-5ec3f1337057 | error  | snapshot1 | 1    |
| f3da162b-5c51-425b-afa8-874a14928ec6 | e84c299d-fafd-4c4f-9a23-5ec3f1337057 | error  | snap2     | 1    |

I've since added another volume/snapshots none work. 
Shouldn't I see some errors on volume log (Debug is set to true) ?

Comment 15 Tzach Shefi 2017-03-07 12:21:56 UTC
Created attachment 1260780 [details]
Packstack answer file ocata.conf, cinder.conf, cinder logs

Forgot to add before, logs files of my system

Comment 16 Eric Harney 2017-03-07 15:32:03 UTC
You need to set nfs_snapshot_support=True in the [nfs] backend section of cinder.conf instead of in [DEFAULT].

Comment 17 Tzach Shefi 2017-03-07 16:34:58 UTC
You do have a good point, sorry my bad. 
So i fixed it restarted services. 

nfs_snapshot_support=True     ( #out the one I touched om  default section)

Now hitting something else, can't even create a volume

2017-03-07 18:25:16.772 17511 ERROR cinder.volume.drivers.nfs [req-44f4716d-9400-413f-997a-b26c737b1da6 - - - - -] Snapshots are not supported with nas_secure_file_operations enabled ('true' or 'auto'). Please set it to 'false' if you intend to  have it enabled.
2017-03-07 18:25:16.772 17511 ERROR cinder.volume.manager [req-44f4716d-9400-413f-997a-b26c737b1da6 - - - - -] Failed to initialize driver.
2017-03-07 18:25:16.772 17511 ERROR cinder.volume.manager Traceback (most recent call last):
2017-03-07 18:25:16.772 17511 ERROR cinder.volume.manager   File "/usr/lib/python2.7/site-packages/cinder/volume/manager.py", line 431, in init_host
2017-03-07 18:25:16.772 17511 ERROR cinder.volume.manager     self.driver.do_setup(ctxt)
2017-03-07 18:25:16.772 17511 ERROR cinder.volume.manager   File "/usr/lib/python2.7/site-packages/cinder/volume/drivers/nfs.py", line 199, in do_setup
2017-03-07 18:25:16.772 17511 ERROR cinder.volume.manager     self._check_snapshot_support(setup_checking=True)
2017-03-07 18:25:16.772 17511 ERROR cinder.volume.manager   File "/usr/lib/python2.7/site-packages/cinder/volume/drivers/nfs.py", line 551, in _check_snapshot_support
2017-03-07 18:25:16.772 17511 ERROR cinder.volume.manager     raise exception.VolumeDriverException(message=msg)
2017-03-07 18:25:16.772 17511 ERROR cinder.volume.manager VolumeDriverException: Volume driver reported an error: Snapshots are not supported with nas_secure_file_operations enabled ('true' or 'auto'). Please set it to 'false' if you intend to  have it enabled.
2017-03-07 18:25:16.772 17511 ERROR cinder.volume.manager 
2017-03-07 18:25:16.773 17511 DEBUG cinder.service [req-44f4716d-9400-413f-997a-b26c737b1da6 - - - - -] Creating RPC server for service cinder-volume start /usr/lib/python2.7/site-packages/cinder/service.py:243
2017-03-07 18:25:16.776 17511 DEBUG oslo_db.sqlalchemy.engines [req-44f4716d-9400-413f-997a-b26c737b1da6 - - - - -] MySQL server mode set to STRICT_TRANS_TABLES,STRICT_ALL_TABLES,NO_ZERO_IN_DATE,NO_ZERO_DATE,ERROR_FOR_DIVISION_BY_ZERO,TRADITIONAL,NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION _check_effective_sql_mode /usr/lib/python2.7/site-packages/oslo_db/sqlalchemy/engines.py:261
2017-03-07 18:25:16.780 17511 DEBUG cinder.service [req-44f4716d-9400-413f-997a-b26c737b1da6 - - - - -] Pinning object versions for RPC server serializer to 1.21 start /usr/lib/python2.7/site-packages/cinder/service.py:250
2017-03-07 18:25:16.780 17511 WARNING py.warnings [req-44f4716d-9400-413f-997a-b26c737b1da6 - - - - -] /usr/lib/python2.7/site-packages/oslo_messaging/rpc/server.py:200: FutureWarning: The access_policy argument is changing its default value to <class 'oslo_messaging.rpc.dispatcher.DefaultRPCAccessPolicy'> in version '?', please update the code to explicitly set None as the value: access_policy defaults to LegacyRPCAccessPolicy which exposes private methods. Explicitly set access_policy to DefaultRPCAccessPolicy or ExplicitRPCAccessPolicy.

2017-03-07 18:25:16.810 17511 INFO cinder.volume.manager [req-44f4716d-9400-413f-997a-b26c737b1da6 - - - - -] Initializing RPC dependent components of volume driver NfsDriver (1.4.0)
2017-03-07 18:25:16.811 17511 ERROR cinder.utils [req-44f4716d-9400-413f-997a-b26c737b1da6 - - - - -] Volume driver NfsDriver not initialized
2017-03-07 18:25:16.811 17511 ERROR cinder.volume.manager [req-44f4716d-9400-413f-997a-b26c737b1da6 - - - - -] Cannot complete RPC initialization because driver isn't initialized properly.
2017-03-07 18:25:26.812 17511 ERROR cinder.service [-] Manager for service cinder-volume panther13.qa.lab.tlv.redhat.com@nfs is reporting problems, not sending heartbeat. Service will appear "down".

I'm guessing this isn't good
 Snapshots are not supported with nas_secure_file_operations enabled ('true' or 'auto'). Please set it to 'false' if you intend to  have it enabled.

While looking at cinder.conf 
Also noticed this one #nas_secure_file_permissions = auto

Guess needs a bit more fiddling around  
Thanks Eric.

Comment 18 Eric Harney 2017-03-07 16:39:49 UTC
(In reply to Tzach Shefi from comment #17)

> While looking at cinder.conf 
> nas_secure_file_operations 
> Also noticed this one #nas_secure_file_permissions = auto
> Guess needs a bit more fiddling around  
> Thanks Eric.

Correct, you'll have to set "nas_secure_file_permissions = False" for this feature to work.

Comment 19 Tzach Shefi 2017-03-15 08:36:23 UTC
Verified on version:

This what I had set on cinder.conf:

nas_secure_file_operations=False  (without which driver fails, default -> true)

Run this twice on a packstack as well as OPSD deployment. 
There is an open OSPD NFS bug bz1416356, had to configure NFS manually post deploy. 

Tempest filtered out 60 snapshot related tests, 47 of them passed. 
Failing tests are same ones for v1 v2, they failed cause I had no connectivity to instance which was expected. 

Tempest summery: 
Ran 60 (+47) tests in 639.747s (+598.497s)
FAILED (id=2, failures=7 (+7), skips=3)

FAIL: tempest.api.volume.test_volumes_snapshots.VolumesV1SnapshotTestJSON.test_snapshot_create_offline_delete_online[compute,id-5210a1de-85a0-11e6-bb21-641c676a5d61]
FAIL: tempest.api.volume.test_volumes_snapshots.VolumesV2SnapshotTestJSON.test_snapshot_create_offline_delete_online[compute,id-5210a1de-85a0-11e6-bb21-641c676a5d61]

FAIL: tempest.api.volume.test_volumes_snapshots.VolumesV1SnapshotTestJSON.test_snapshot_create_with_volume_in_use[compute,id-b467b54c-07a4-446d-a1cf-651dedcc3ff1]
FAIL: tempest.api.volume.test_volumes_snapshots.VolumesV2SnapshotTestJSON.test_snapshot_create_with_volume_in_use[compute,id-b467b54c-07a4-446d-a1cf-651dedcc3ff1]

FAIL: tempest.api.volume.test_volumes_snapshots.VolumesV1SnapshotTestJSON.test_snapshot_delete_with_volume_in_use[compute,id-8567b54c-4455-446d-a1cf-651ddeaa3ff2]
FAIL: tempest.api.volume.test_volumes_snapshots.VolumesV2SnapshotTestJSON.test_snapshot_delete_with_volume_in_use[compute,id-8567b54c-4455-446d-a1cf-651ddeaa3ff2]

FAIL: tempest.scenario.test_snapshot_pattern.TestSnapshotPattern.test_snapshot_pattern[compute,id-608e604b-1d63-4a82-8e3e-91bc665c90b4,image,network]

Comment 22 errata-xmlrpc 2017-05-17 19:22:10 UTC
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.