📝 devlog v5 [ 2025-12-06 ]

- restoring microcluster to LAN
- microcluster - cm5 4gb ram 128gb nvme (native) + wyse 3040 + wyse 3040
- all attempts connecting to main network failed
- constant no route to host errors
- solved by power cycling office switch, powering down all nodes and poe switch
- set static ip's with routes to router ip (192.168.0.1)
- powering on office switch, cm5, then poe switch which powers the wyse 3040 nodes & luckfox sbc.
- reserved addresses in router
- grafana works perfectly over wifi, using laptop.
- kept cron job that pings via bash script to all nodes to prevent dhcp drift

🔍 insights

- wyse nodes have minimal on-board storage solved by using usb to nvme adapters and external nvme drives (128gb)
- no difference between using sata hdd/sdd, as bottlenecked by usb 2.0 speeds
- nvme has a small form factor and fits better on the microcluster.
- i believe the issue stems from stale arp tables on the office switch.
- all solved and can now manage microcluster from any device on the network.