Roadmap · cluster hardware
Where the hardware is going
The walk-through only describes what is running today. This page lists the physical side of what comes next: the two nodes being added, the memory and cooling going into the existing machines, the storage that replaces the 2012 disks, and the links that need to be sturdier. Software, services, and side projects are deliberately left off. It reflects the plan as of late October 2026.
In progressNext upPlannedWaiting on something
Compute and GPUs
- In progressNode 5: a secondhand Dell workstationA refurbished tower workstation bought in October, a 4-core Xeon with 64 GB of ECC memory and two full-length x16 slots, was originally a loaner for testing the new GPUs and was going back to the seller. It is being kept instead and joins the cluster as the fifth node: a boot drive, a Proxmox install, the firewall allow-list, metrics, the thermal guard, and the nightly backup job. It still hosts the day-one GPU test before the cards move to their final home.
- In progressThe open-frame AI node, node 6A Threadripper PRO 3945WX on a workstation board that was already on hand, 128 GB of ECC memory in eight sticks, two used RTX 3060 12 GB cards, a big tower cooler, and a 1000 W supply on an open frame. Every part has been bought. The cards are power-capped at 150 W each.
- In progressMemory upgrades for the three desktopsSeven 16 GB sticks are on order to take the desktops to 32, 64, and 64 GB. Node 1's kit is rated for 3200 MT/s but runs at 2133 until its memory profile is switched on in firmware; that is a reboot and a to-do.
- PlannedRetire the five Maxwell cardsOnce the AI node is stable and has carried the inference load for a week of overlap, the M2000s come out of the desktops and are sold.
Cooling and power
- PlannedCoolingA tower cooler for node 1, a cleaner cooler for the desktop that duty-cycles at 82 °C under load, and better airflow for the room the whole thing sits in.
- WaitingBigger battery backupThe current UPS idles around 60% of capacity, node 5 will push that higher, and the AI node cannot go on it at all. A larger unit, or a second one dedicated to the AI node, comes before that node runs unattended.
- Next upPower-on after an outageCheck every node's firmware restores power after AC loss, and set the ones that don't. Until then a power cut needs a human to press five buttons in the right order.
Storage
- Next upFinish the disk swapThe failing 2012 drive is detached from the pool but still physically in its bay; it comes out, the SAS cable on the new 4 TB drive gets reseated, and a vibration fix goes in. Then decide: keep a hot spare, or start replacing the other four with 4 TB drives so the pool can finally grow.
- PlannedA 10 TB replacement NAS and a 10 TB offsite twinNothing is offsite today. The plan is two matching builds: one at home to replace the 2012 array, one at a second location receiving nightly replication over the mesh VPN. That covers the backups, the media library, and the model archive in one move.
Network
- PlannedA second cluster heartbeat linkThe cluster's quorum traffic runs on the single LAN. A second ring on a separate switch and cable run stops a busy or flapping switch from making the cluster fence itself.
- WaitingA better uplinkThe satellite link has a five-minute renewal flap and evening congestion. A free dish upgrade has been offered and not yet accepted; a fibre build is being registered for; a local wireless ISP is the fallback.
Items leave this page when they ship and appear on the walk-through instead. Dates are deliberately rough: this is a hobby with a day job in front of it.