From 7ffde21559652ccaabb198dcf975b2a6c1cb04c1 Mon Sep 17 00:00:00 2001 From: Csaba Henk Date: Thu, 17 Dec 2015 01:46:24 +0100 Subject: [PATCH] glusterfs: document Gluster NFS misbehavior Gluster NFS does not disrupt pre-existing mounts on access revocation. Change-Id: Ie0a10a495acf78744bd534fc7ebb190c242036c0 Partial-Bug: #1506659 --- doc/source/devref/glusterfs_driver.rst | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/doc/source/devref/glusterfs_driver.rst b/doc/source/devref/glusterfs_driver.rst index 3550abfa..f69ca3cf 100644 --- a/doc/source/devref/glusterfs_driver.rst +++ b/doc/source/devref/glusterfs_driver.rst @@ -181,6 +181,13 @@ Known Restrictions currently created within a single GlusterFS volume of a GlusterFS storage pool. - The driver does not provide read-only access level for shares. +- Assume that share S is exported through Gluster NFS, and tenant machine T + has mounted S. If at this point access of T to S is revoked through + `access-deny`, the pre-existing mount will be still usable and T will still + be able to access the data in S as long as that mount is in place. + (This violates the principle *Access deny should always result + in immediate loss of access to the share*, see + http://lists.openstack.org/pipermail/openstack-dev/2015-July/069109.html.) The :mod:`manila.share.drivers.glusterfs` Module ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~