[cinder] ceph multiattach details?

Martinx - ジェームズ thiagocmartinsc at gmail.com
Tue May 28 19:07:26 UTC 2019

Last time I checked OpenStack Ansible, Manila wasn't there... I believe
that they added support for it in Stein but, I'm not sure if it supports
CephFS backend (and the required Ceph Metadata Containers, since CephFS
needs if, I believe).

I'll definitely give it a try!

Initially, I was planning to multi-attach an RDB block device against 2 or
more Instances and run OCFS2 on top of it but, Manila with CephFS looks way


On Mon, 27 May 2019 at 22:30, Erik McCormick <emccormick at cirrusseven.com>

> On Mon, May 27, 2019, 5:33 PM Martinx - ジェームズ <thiagocmartinsc at gmail.com>
> wrote:
>> Hello,
>>  I'm very curious about this as well!
>>  It would be awesome to support Cinder multi-attach when using Ceph... If
>> the code is already there, how to use it?!
>> Cheers,
>> Thiago
>> On Mon, 27 May 2019 at 03:52, Bernd Bausch <berndbausch at gmail.com> wrote:
>>> The Stein release notes mention that the RBD driver now supports
>>> multiattach, but i have not found any details. Are there limitations? Is
>>> there a need to configure anything?
>>> In the RBD driver
>>> <https://opendev.org/openstack/cinder/src/branch/master/cinder/volume/drivers/rbd.py#L767>,
>>> I find this:
>>>     def _enable_multiattach(self, volume):
>>>         multipath_feature_exclusions = [
>>>             self.rbd.RBD_FEATURE_JOURNALING,
>>>             self.rbd.RBD_FEATURE_FAST_DIFF,
>>>             self.rbd.RBD_FEATURE_OBJECT_MAP,
>>>             self.rbd.RBD_FEATURE_EXCLUSIVE_LOCK,
>>>         ]
>>> This seems to mean that journaling and other features (to me, it's not
>>> quite clear what they are) will be automatically disabled when switching on
>>> multiattachment.
>>> Further down in the code I see that replication and multiattach are
>>> mutually exclusive.
>>> Is there some documentation about the Ceph multiattach feature, even an
>>> email thread?
>>> Thanks,
>>> Bernd
> There isn't really a Ceph multi-attach feature using Cinder. The code
> comment is stating that, while the Openstack side of things is in place,
> Ceph doesn't yet support it with RBD due to replication issues with
> multiple clients. The Ceph community is aware of it, but has thus far
> focused on CephFS as the shared file system instead.
> This could possibly be used with the NFS Cinder driver talking to Ganesha
> with CephFS mounted. You may also want to look at Openstack's Manilla
> project to orchestrate that.
> -Erik
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-discuss/attachments/20190528/50d788c6/attachment.html>

More information about the openstack-discuss mailing list