I'm in the process of designing a new HA cluster across two locations. The two locations are connected together with multiple dark fibers and I would like to create an robust HA system that can migrate vm's between the locations.
For storage I'm planning on using Ceph as I have experience with it and within Ceph you can define failure domains so when one of the locations fails the whole system can run on the other location.
As I want to have 3 nodes (to have an quorum) on each location with an total of 6 nodes, which is an even number, what is bad. I was wondering if there is something like a failure domain in Proxmox HA/Pacemaker?
Or how do other people have an similar kind of setup? As I said, this is currently all theory and no hardware has been purchased at this moment so I'm open to suggestions.
Thank you in advance!
For storage I'm planning on using Ceph as I have experience with it and within Ceph you can define failure domains so when one of the locations fails the whole system can run on the other location.
As I want to have 3 nodes (to have an quorum) on each location with an total of 6 nodes, which is an even number, what is bad. I was wondering if there is something like a failure domain in Proxmox HA/Pacemaker?
Or how do other people have an similar kind of setup? As I said, this is currently all theory and no hardware has been purchased at this moment so I'm open to suggestions.
Thank you in advance!