Thank you Ruben, I did that way you just described earlier this week this new setup. My environment will grow and i am have some particularities on it as well, even the storage thing.
I am doing some customizations for my needs pontually, but also I have found some people in the list that faecd some issues like mine regarding an OpenFlow based network. Again, thanks for the heads up, it was exactly what I got, was tyring to avoid some erroneous setup for a near future. Cheers. On Thu, Oct 4, 2012 at 6:07 PM, Ruben S. Montero <[email protected]>wrote: > Hi Rodolfo > > Yes you email was posted in the right place ;) > > Some thoughts, inlined.... > > > > > My main concern is, is it possible to have multiple datastore types > inside > > the cluster (like adding future nodes to act as storage or any other > NAS/SAN > > device?) > > Is there a good way to use this set up looking for future upgrades in > terms > > of resources? > > Yes it is possible to have multiple storage backends, that was the > main reason to introduce the datastore abstraction. > > In order to ease the upgrade to add more datastores (even of the same > type, e.g. you buy a new NAS server and want to use both NFS > datastores), I'd recommend not to mount/export the whole > var/lib/one/datastore/ directory, but just the each datastore > directory individually. > > For example if you have an image datastore (e.g. 100), export > /var/lib/one/datastore/100 in your NFS server (that directory will be > the datastore storage area) and mount that under > /var/lib/one/datastore/100 in the hosts. For 3.6, do not mount it > elsewhere and link it there (there is a bug solved in 3.8). And the > same for the system datastore 0 (take a look at this picture [1]). > Note that the system datastore only needs to be added to the hosts > that run VMs. > > [1] > http://opennebula.org/_detail/documentation:rel3.4:fs_shared.png?id=documentation%3Arel3.6%3Afs_ds > > > > > *Other point, the other reason i am asking in which one to use is: Is it > > possible to extend the created LVs on the fly without compromising the > > running instance? > > Regarding that performance is not a big concern right now. > > This would need some hacking, Take a look at > > > http://lists.opennebula.org/pipermail/users-opennebula.org/2012-September/020357.html > > > > > I hope I was a little bit concise in here. :) > > I am stuck in a crossroad in which datastore type to use and any advice, > > insight or tip would be very welcome and very helpful! > > For your simple setup I'd recommend to go for the FS datastore, using > NFS. The details in the documentation should be enough to set up this > [2,3], but come back if you need more help > > [2] System datastore: http://opennebula.org/documentation:rel3.6:system_ds > [3] FS Datastore: http://opennebula.org/documentation:rel3.6:fs_ds > > > Good Luck > > Best > > Ruben > > > > Thanks in advance! > > > > Cheers, > > > > -- > > Have you tried turning it off and on again? > > > > _______________________________________________ > > Users mailing list > > [email protected] > > http://lists.opennebula.org/listinfo.cgi/users-opennebula.org > > > > -- > Ruben S. Montero, PhD > Project co-Lead and Chief Architect > OpenNebula - The Open Source Solution for Data Center Virtualization > www.OpenNebula.org | [email protected] | @OpenNebula > -- Have you tried turning it off and on again?
_______________________________________________ Users mailing list [email protected] http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
