01 / THE ROLE
In this lab.
The compute foundation is a three-node Proxmox VE cluster, using a mix of compact AMD systems and a larger HP server. Linux guests provide separate environments for application hosting, databases and supporting infrastructure. This lets me maintain a service environment independently of the physical machine beneath it.
02 / THE CONNECTIONS
Follow the flow.
- STEP 01Physical hosts
- STEP 02Proxmox cluster
- STEP 03Virtual machines
- STEP 04Services
03 / How it works
How it works.
A service starts with a guest sized for its purpose. The hypervisor manages virtual CPU, memory, disks and virtual network interfaces; the guest runs its own operating system and applications. Selected workloads use Proxmox HA and ZFS replication. Those are workload-specific arrangements, not a promise that every VM can move freely to every node.
04 / Persistent data & recovery
Persistent data & recovery.
Guest configuration, virtual disks and the state inside the guest are separate parts of recovery. Proxmox Backup Server provides a VM-backup layer, while application recovery still needs an understanding of databases, files and credentials. Replication supports availability; it is not an independent backup.
05 / How to check it
How to check it.
Check cluster membership and quorum, host resources, storage health, guest availability and the actual application. A green cluster does not prove that a database has started or that a web application works. Following a migration or failover, validate dependencies inside the guest.
06 / Decisions & limits
Decisions & limits.
A heterogeneous cluster is useful for learning and reusing hardware, but available capacity and storage placement constrain failover. I treat HA as a specific design to validate rather than a blanket property of the whole lab.