On 03/06/14 02:24, Mark Nelson wrote:
> It's kind of a tough call. Your observations regarding the downsides of
> using NFS with RBD are apt. You could try throwing another distributed
> storage system on top of RBD and use Ceph for the replication/etc, but
> that's not really ideal either. CephFS is relatively stable with
> active/standby MDS configurations, but it may still have bugs and there
> are no guarantees or official support (yet!).
The big thorn in the side here is when people on Windows clients want to
effectively access CephFS. I'm not sure if there's a Win32 port of FUSE
that cephfs-fuse could be ported to. So until then you're stuck with
trying to cluster Samba or putting up with a bottleneck/single
point-of-failure.
I understand that for CephFS: the officially supported configuration is
a single MDS where as MDS HA is a work in progress. At the time when I
was setting up our Ceph cluster, CephFS was considered experimental even
though in testing it seemed to be working fine, so I left that bit out
of the production cluster, figuring I'd add it later.
If I deploy an MDS, and that machine goes down, is it
possible/sufficient to deploy a MDS to another node to get back up and
running without data loss?
--
Stuart Longland
Systems Engineer
_ ___
\ /|_) | T: +61 7 3535 9619
\/ | \ | 38b Douglas Street F: +61 7 3535 9699
SYSTEMS Milton QLD 4064 http://www.vrt.com.au
_______________________________________________
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
