My reasons for doing it this way is that it will be highly resilient for our business and will be a lot easier to scale up as we need it to. This is an overkill setup for what we are running currently but I can see us becoming more dependent on self hosting as it is cheaper than using the cloud for random things we need to run. I have already spun up servers for monitoring, helpdesk, and backups. All would have not existed had I not had some old servers to play around with.
Virtual server is currently going to be a single self-hosted oVirt install for VM use. So in essence the whole thing will look something like this:
Storage servers will be set out as such: CentOS 7 for OS, md RAID1 for 8 drives (4 groups of 2, 2TB each drive), LVM PV for all RAID blocks, GlusterFS Bricks are contained within a LV. This config will span 2 servers.
My idea is to deploy an oVirt cluster with a GlusterFS cluster as a storage backend, servers are connected via dual 10GB NICs to talk between the clusters and a normal 1GB NIC is used as management and external access for the clustered services.
Is anyone able to help me sanity check some infrastructure that I would like to deploy at work? We are in the process of moving away from an SBS2011 install and a very small VMWare installation (only 2 VMs currently). I would like to put in a virtualisation platform that would allow us to use more services as our needs grow. I would also like to use as much OSS/FOSS stuff as possible without making it too complicated. Any opinions or recommendations are welcome.
System Administrator working in the Logistics Sector.
I enjoy most things GNU, Linux, Music, Art, and Games.
This Mastodon instance is for people interested in technology. Discussions aren't limited to technology, because tech folks shouldn't be limited to technology either!