Skip to main content

Common workflows

Short operator paths for the jobs you run most often. Open the console at https://<host>:9095/app after signing in (see Getting Started). Never publish lab IPs — use <host> or 127.0.0.1.

Workflow index​

WorkflowRouteGuide
Create a VM/app/createCreate VM
Manage the fleet/app/vmsVirtual Machines
Per-VM Network Fabric (schema v4)/app/vms/:name → DataplaneVM Dataplane
Edge Maglev / Service Fabric v6/app/edge-dataplane → ServicesEdge Dataplane
Container Groups (Secure Containers)/app/container-groupsContainer Groups
Agents / Sessions/app/agents, /app/sessionsAgents · Sessions
Snapshots/app/snapshots or VM → SnapshotsSnapshots
Backups & restore/app/backupsBackups

Create a VM​

  1. Open Core → Create VM (/app/create).
  2. Basics — name + catalog image (or host path).
  3. Resources — vCPUs, memory, disk. Networking:
    • NAT (default) — private outbound; add Expose ports / Expose SSH (22) for inbound from clients that can reach the host.
    • Bridged — own LAN address (DHCP or static via cloud-init). Use bridged when you will attach VM Dataplane eBPF.
  4. Review → create → land on /app/vms/:name.
  5. Empty / fail: Image missing or FluxVM unreachable — check Dashboard capability chips and Admin basics.
  6. Success: VM appears on /app/vms with the expected state badge.

Attach VM Dataplane (Network Fabric schema v4)​

  1. Create or pick a bridged VM with a host TAP (network_tap).
  2. Open /app/vms/:name → Dataplane (or Network → Open Dataplane).
  3. On Status, confirm Attached = yes, Mode = ebpf, Schema version = 4.
  4. On Policy, pick a preset or edit allow/deny CIDRs and tcp/PORT / udp/PORT ports → Save policy.
  5. Use Effective, Stats, and Flows to verify merges and allow/drop counters.
  6. Empty / fail: Schema missing or not attached — enable FluxVM [sandbox.dataplane] mode = "ebpf" and confirm the TAP exists (operator guide linked from the dataplane page).
  7. Success: Status shows schema 4 attached; counters move when traffic matches policy.

Create an Edge Maglev service (Service Fabric v6)​

  1. Open Infrastructure → Edge Dataplane (/app/edge-dataplane).
  2. Open the Services tab (Service Fabric v6 / BPF schema 4).
  3. Upsert a Maglev VIP service (backends, affinity, optional drain/health, pacing / host-routing fields the form exposes).
  4. Check Health for BPF/bpffs presence; use ads / conntrack / HA controls as needed.
  5. Empty / fail: Health not ok — FluxVM service dataplane not enabled or bpffs missing; see Edge Dataplane and ebpf-service-fabric.md.
  6. Success: Service row appears; schema badge shows Service Fabric v6; backend health reconciles.

Take and revert a snapshot​

  1. Prefer Virtual Machines → open the VM → Snapshots, or Operations → Snapshots (/app/snapshots) and enter the VM name.
  2. Create — name, optional description, type Disk (default, fast) or Full (disk + memory, slower).
  3. To roll back: stop the VM, then Revert on the snapshot row (confirmation required).
  4. Empty / fail: Create 409 right after start — QMP not ready yet; wait and retry. Revert while running is rejected.
  5. Success: Snapshot listed with type and timestamp; after revert, VM boots from that disk state.

Full notes: Snapshots.

Backup and restore​

  1. Open Operations → Backups (/app/backups).
  2. Create Backup — pick VM, Full or Incremental (compressed, 30-day retention by default).
  3. Watch Active Jobs until status is completed.
  4. Restore — in place on the original VM, or Restore to new VM with a new name (config + disks; not live runtime state).
  5. Empty / fail: Job stuck/failed — check Active Jobs error text, storage pool space, and FluxVM health.
  6. Success: Backup row shows completed with size/expiry; restore produces a runnable VM on /app/vms.

Guide: Backups. Recurring jobs: Backup Scheduler.