I'm rebuilding my cluster and I'd like to take a stab at Ceph. Here's what I Have to work with:
6 servers, all dual-xeons with at least 6 cores (24 threads). min 128GB ram.
Disks are where things get weird:
My plan is to link everything up with 25gb dedicated NICs for the ceph network.
Is this unreasonable?
6 servers, all dual-xeons with at least 6 cores (24 threads). min 128GB ram.
Disks are where things get weird:
- 7 x 3TB platter (21tb)
- 5 x 2TB platter (10tb)
- 5 x 2TB platter (10tb)
- 1 500GB platter + 8 x 1.5 TB nvme (12tb)
- 7 x 3TB nvme (21tb)
- 4 x 750GB nvme (3 TB) with room to grow.
My plan is to link everything up with 25gb dedicated NICs for the ceph network.
Is this unreasonable?