You can survive loosing 2 servers out of 5 by using disperse volumes with 3+2 configuration. But in your case, web hosting, with lots of small files and random read/write it is not recommended. Maybe you can test the workload and give it a try. Other than disperse volumes I don't know a solution that will handle failure of 2 out of 5 servers. On Tue, Feb 16, 2016 at 11:57 PM, Thorvald Hallvardsson <thorvald.hallvardsson@xxxxxxxxx> wrote: > Hi guys, > > I need to build a backend storage for the website hosting. Storage needs to > be highly available and easily expandable. I have 5 servers with 300GB disks > for that purpose. I would like to create volume replicated with a 100GB of > size. I want the configuration which will accept 2 boxes being down at the > same time without losing any data. > > Can you please suggest what is the desired configuration of the brick in > that case ? > > I have seen some posts here on the group where people mentioned something > that some replica configurations works in expected - from gluster point of > view - way but unexpected - from the user point of view - where you lose one > box and the volume goes RO. I want to avoid that situation. > > I also want to use FUSE driver and configure it in the way that clients will > be able to access all 5 servers at the same time. > > Thank you in advance for your help. > > Regards, > TH > > _______________________________________________ > Gluster-users mailing list > Gluster-users@xxxxxxxxxxx > http://www.gluster.org/mailman/listinfo/gluster-users _______________________________________________ Gluster-users mailing list Gluster-users@xxxxxxxxxxx http://www.gluster.org/mailman/listinfo/gluster-users
