Hi.
As we would like to use the CEPH storage with CloudStack/XS we have to
use NFS or iSCSI client nodes to provide shared storage. To avoid having
several nodes of physical hardware we thought that we could run
NFS/iSCSI client node on the same box with Ceph OSD node. Possibly we
could even run ceparate VMs for MONs on the same hypervisor. This would
give us flexibility and we could migrate NFS/iSCSI or MONs VMs to any
hosts we want any time.
Also we could take snapshot of the Ceph OSD VMs during upgrades. If
something does not go well, we can roll back fast.
Potentially, we could turn every XS local storage to Ceph OSD (utilize
the unused local HDDs).
I think the I/O performance tax from VM to raw local disk is negligible
in comparison with physical box.
Anyway, these are just thoughts. Might not be the best idea, but that is
the reason why I am asking.
Thx Jiri
On 21/08/2015 10:12, Steven McDonald wrote:
Hi Jiri,
On Thu, 20 Aug 2015 11:55:55 +1000
Jiri Kanicky <j@xxxxxxxxxx> wrote:
We are experimenting with an idea to run OSD nodes in XenServer VMs.
We believe this could provide better flexibility, backups for the
nodes etc.
Could you expand on this? As written, it seems like a bad idea to me,
just because you'd be adding complexity for no gain. Can you explain,
for instance, why you think it would enable better flexibility, or why
it would help with backups?
What is it that you intend to back up? Backing up the OS on a storage
node should never be necessary, since it should be recreatable from
config management, and backing up data on the OSDs is best done on a
per-pool basis because the requirements are going to differ by pool and
not by OSD.
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
