Physical Network IPs

• 192.168.178.1 - Router + DHCP Server
• 192.168.178.10 - 192.168.178.20 -> Servers
• 192.168.178.100-255 -> DHCP

Virtual Network IP

• 192.168.179.1 - 192.168.179.10 -> Servers VIP
  • 192.168.179.1 - K3S server group keepalived
    • Alfheimr
    • Midgard
    • Niflheimr
    • Jotunheimr
• 192.168.179.11 - 192.168.179.255 - Virtual IPs (MetalLB)

HW

• HP T620T (Alfheimr)
  • IP: 192.168.178.10
  • Roles:
    • Keepalived
    • Worker
• ThinkCentre M710q (Midgard)
  • 192.168.179.11
  • Roles:
    • K3S Master 1
    • Keepalived
• Firebat AM02 (Niflheimr)
  • 192.168.179.12
  • Roles:
    • K3S Master 2
    • Keepalived
• SOYO M2PLUS (Jotunheimr)
  • 192.168.179.13
  • Roles:
    • K3S Master 3
    • Keepalived

Physical Hardware Placement

Server Rack:

• older router
• Alfheimr
• Midgard
• Niflheimr
• Jotunheimr

TBD:

• NAS

Infrastructure

• Keepalived
• RClone mount to seedbox ftp on each node to /mnt/seedbox
• K3S
  • ArgoCD
  • MetalLB
  • CertMgr ( *.k8s.kirillorlov.pro )
  • Longhorn (backup to OVH S3)
  • Cloudflared
  • prometheus-operator (node-exporter, grafana, prometheus)
  • External DNS (to sync with pi-hole)
  • PiHole as DNS server
    • 2 installations - PR and DR
    • OrbitalSync to sync from PR to DR
  • External Secret Operator
  • Reloader
  • MariaDB-operator
  • VerticalPodAutoscaler

Apps

Running:

• rss-to-telegram
• nextcloud (+mariadb)
• hugo (website)
• photoprism
• actualbudget
• paperless (+redis)
• jackett
• sonarr
• radarr
• plex
• overseerr
• bonsai
• homepage

Restore cluster

• create new cluster
• add argocd
• add eso and longhorn
• restore all volumes from longhorn
• create pv and pvc (using button in longhorn ui)
• add meta-app - it will reconfigure all back

ToDo how to:

• Setup proper liveness / startup / readyness checks for userapps
• Ensure all tags are not current - so that renovate would work
• HA for ALL apps
• RSSTT - switch to statefulset
• All ingresses - check http(s)
  • or even better - write a policy for Kyverno for that