Re: [ceph-users] Multiple cephfs filesystems per cluster
Thanks John - It did look like it was heading in that direction! I did wonder if a 'fs map' & 'fs unmap' would be useful too; filesystem backups, migrations between clusters & async DR could be facilitated by moving underlying pool objects around between clusters. Dave On Wed, Sep 17, 2014 at 11:22 AM, John Spray wrote: > Hi David, > > We haven't written any code for the multiple filesystems feature so > far, but the new "fs new"/"fs rm"/"fs ls" management commands were > designed with this in mind -- currently only supporting one > filesystem, but to allow slotting in the multiple filesystems feature > without too much disruption. There is some design work to be done as > well, such as how the system should handle standby MDSs (assigning to > a particular filesystem, floating between filesystems, etc). > > Cheers, > John > > On Wed, Sep 17, 2014 at 11:11 AM, David Barker > wrote: > > Hi Cephalopods, > > > > Browsing the list archives, I know this has come up before, but I thought > > I'd check in for an update. > > > > I'm in an environment where it would be useful to run a file system per > > department in a single cluster (or at a pinch enforcing some client / fs > > tree security). Has there been much progress recently? > > > > Many thanks, > > > > Dave > > > > > > > > ___ > > ceph-users mailing list > > ceph-users@lists.ceph.com > > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > > > ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] Multiple cephfs filesystems per cluster
Hi David, We haven't written any code for the multiple filesystems feature so far, but the new "fs new"/"fs rm"/"fs ls" management commands were designed with this in mind -- currently only supporting one filesystem, but to allow slotting in the multiple filesystems feature without too much disruption. There is some design work to be done as well, such as how the system should handle standby MDSs (assigning to a particular filesystem, floating between filesystems, etc). Cheers, John On Wed, Sep 17, 2014 at 11:11 AM, David Barker wrote: > Hi Cephalopods, > > Browsing the list archives, I know this has come up before, but I thought > I'd check in for an update. > > I'm in an environment where it would be useful to run a file system per > department in a single cluster (or at a pinch enforcing some client / fs > tree security). Has there been much progress recently? > > Many thanks, > > Dave > > > > ___ > ceph-users mailing list > ceph-users@lists.ceph.com > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] Multiple cephfs filesystems per cluster
On 09/17/2014 12:11 PM, David Barker wrote: > Hi Cephalopods, > > Browsing the list archives, I know this has come up before, but I thought > I'd check in for an update. > > I'm in an environment where it would be useful to run a file system per > department in a single cluster (or at a pinch enforcing some client / fs > tree security). Has there been much progress recently? > No, that's not possible. It's a single hierarchy. However, you can create subdirectories per department and do a subtree mount. ACLs and tree security isn't implemented yet however. > Many thanks, > > Dave > > > > ___ > ceph-users mailing list > ceph-users@lists.ceph.com > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > -- Wido den Hollander 42on B.V. Ceph trainer and consultant Phone: +31 (0)20 700 9902 Skype: contact42on ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] Multiple CephFS filesystems per cluster
> Hi, > > Is it possible to have more than one CephFS filesystem per Ceph cluster? > > In the default configuration, a ceph cluster has got only one filesystem, and > you can mount that or nothing. Is it possible somehow to have several > distinct > filesystems per cluster, preferably with access controls that could make ist > possible for one user to access only one filesystem and not the other? > Maybe not exactly the answer you are looking for, but you can pass the -r option to ceph-fuse and that will mount that directory as the root of the fs, so in my ceph fs I have a directory called /installmedia, and I mount that on /var/installmedia, so I do: ceph-fuse -r /installmedia /var/installmedia James ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] Multiple CephFS filesystems per cluster
On Wed, Aug 21, 2013 at 11:33 AM, Guido Winkelmann wrote: > Hi, > > Is it possible to have more than one CephFS filesystem per Ceph cluster? > > In the default configuration, a ceph cluster has got only one filesystem, and > you can mount that or nothing. Is it possible somehow to have several distinct > filesystems per cluster, preferably with access controls that could make ist > possible for one user to access only one filesystem and not the other? Right now you can only have a single filesystem per cluster, and there are sadly no controls to forcibly separate users in the tree. You *can* mount only a subtree of the full filesystem, and if you control root on the mounting nodes this might be sufficient. (We've also done some scoping work on securing this for a community blueprint in the Dumpling CDS, but unfortunately it wasn't picked up: http://wiki.ceph.com/01Planning/02Blueprints/Dumpling/Client_Security_for_CephFS.) You can also segregate the file data into different RADOS pools and restrict access to those pools on a per-client basis, but that leaves everybody's metadata visible to and subject to manipulation by the others. -Greg Software Engineer #42 @ http://inktank.com | http://ceph.com ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com