This morning, while I was in the shower, I thought of a workaround for this. Here’s my idea:
- Use hardware such as an HPE Apollo.
- Install a hypervisor like ESXi (make sure to check the compatibility matrices!).
- Optionally, set up a vCenter (I would recommend it).
- Create N virtual machines, each with an appropriate disk quota, which will become Linux hardened repositories. You would create as many VMs as there are Veeam servers that need to use them.
- As per best practices, physical interfaces and virtual switches, etc., should be separated between the VMs' uplink and the ESXi/vCenter Management.
- The Management physical interfaces and those of the ILO or iDRAC should be connected to a small, dedicated switch, isolated from the rest of the network.
- For management, you could connect a fully dedicated NUC/miniPC, isolated from the network and only connected to the small switch (alternatively, you could connect to the switch with your laptop).
Does this make sense, or is it a bad idea? What do you think? Any comments and corrections are appreciated
https://blog.tinivelli.com/one-veeam-ha ... 749d5c4ffe