Created attachment 1322773 [details] cinder.log Description of problem: Cinder backup uses os_brick to create a nfs share mount point as root and fails creating subdirectories later on as cinder-backup is running as cinder and tries to create a sub directory as cinder. Version-Release number of selected component (if applicable): How reproducible: All the time Steps to Reproduce: 1. Take a volume backup with a NFS backend Actual results: We get a permission denied Expected results: Additional info: def _init_backup_repo_path(self): remotefsclient = remotefs_brick.RemoteFsClient( 'nfs', utils.get_root_helper(), nfs_mount_point_base=self.backup_mount_point_base, nfs_mount_options=self.mount_options) remotefsclient.mount(self.backup_share) LOG.info("Mount point: %s", remotefsclient.get_mount_point(self.backup_share)) (out, err) = utils.execute('chown', 'cinder.cinder', remotefsclient.get_mount_point(self.backup_share), run_as_root=True) LOG.info("Result: %s %s", out, err) return remotefsclient.get_mount_point(self.backup_share
Patch has been backported upstream to Pike, next effort is to propose the patch to driverfixes/ocata and driverfixes/newton. Then we can backport to OSP releases.
Verified on: openstack-cinder-9.1.4-40.el7ost.noarch Created a volume, in my case from an image to add some data. #cinder create 2 --image-id cad7c940-27e4-418c-83de-4490986eb59a --name cirrosvol Configure Cinder backup with nfs, restart cinder backup service Backup the new volume cinder backup-create beebd668-2410-4602-9030-2a6f1937150f +-----------+--------------------------------------+ | Property | Value | +-----------+--------------------------------------+ | id | 489b141c-dcf6-4452-940e-07618849c4c5 | | name | None | | volume_id | beebd668-2410-4602-9030-2a6f1937150f | +-----------+--------------------------------------+ Backup available -> cinder backup-list +--------------------------------------+--------------------------------------+-----------+------+------+--------------+--------------------------------------------+ | ID | Volume ID | Status | Name | Size | Object Count | Container | +--------------------------------------+--------------------------------------+-----------+------+------+--------------+--------------------------------------------+ | 489b141c-dcf6-4452-940e-07618849c4c5 | beebd668-2410-4602-9030-2a6f1937150f | available | - | 2 | 3 | 48/9b/489b141c-dcf6-4452-940e-07618849c4c5 | +--------------------------------------+--------------------------------------+-----------+------+------+--------------+--------------------------------------------+ Check controller where backup service is running look for mount, or on nfs server itself, notice a new folder was created. Before backup folder is empty -> [root@cougar11 ins_backup]# ll total 0 After backup is imitated we see a new folder created with cinder:cinder [root@cougar11 ins_backup]# ll total 0 drwxr-xr-x. 3 cinder cinder 16 אוג 7 15:30 48 Works as expected.
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. https://access.redhat.com/errata/RHBA-2018:2717