Console Recovery

Console Recovery #

Getting back into an infrastructure device when the network can no longer reach it, and returning it to declared state afterwards.

Every device in this section is managed through the network it helps provide. The core router’s management path is a rule on the core router; the switch’s management address rides the VLANs the switch defines; the hypervisors reach the world through a bridge they configure themselves. That circularity is normal and mostly invisible — it matters only on the day a change removes the path used to make changes. These pages are what you follow on that day.

By device #

DeviceAddressWay inPage
Core router dv02cor002p0110.20.99.1Mini DisplayPort + USB keyboardCore router
Hypervisors dv02hyp001p01, dv02hyp002p0210.20.99.21, .22DisplayPort→HDMI adapter + USB keyboardHypervisor
Access switch dv02acc001p0110.20.99.10Factory reset, then reapplyAccess switch
Access point dv02wap001p0110.20.99.9Wired into the switch first, then factory resetWireless AP

Two kinds of recovery #

The router and the hypervisors are repaired in place. Both hold state that is expensive or impossible to reconstruct — the router’s configuration history, the hypervisors’ running VMs — so the console exists to fix the one thing that broke and leave the rest alone.

The switch and the AP are reset and reapplied. Their entire configuration is declared in inventory and rebuilt by a playbook, so the fastest route back is usually to discard the device’s state rather than diagnose it. The cost is a bootstrap step: a factory-reset device is not at its inventory address, so automation cannot reach it until you put it back on the network by hand.

The AP adds one more, and it is the one that catches people out: it is the path you would normally use to reach everything else here. Wireless management dies with it, so its recovery begins by cabling into the access switch rather than by touching the AP at all — and the laptop leases an address from the management segment’s DHCP pool, or takes a hand-set one if the pool is gone. Port gigabitEthernet 1/0/2 on the access switch is declared and held empty for this; see Wireless AP step 1.

Keep findable #

ItemNeeded for
Mini DisplayPort cableCore router — unusual connector, only ever needed on the worst day
Patch cable + USB-Ethernet adapterThe AP — it carries the operator’s own connectivity, so recovering it starts by going wired
DisplayPort→HDMI adapterHypervisors
USB keyboardAll console work
A monitor that is not on the affected networkAll console work

Before you need any of this #

Take a config backup before a disruptive change to the router: System → Configuration → Backups → Download, as the segmentation prerequisites already require. OPNsense keeps its own history on the box, which covers a configuration mistake but not a failed disk, and there is no off-box backup automation in any collection.

For the switch and the AP there is nothing to back up — inventory is the backup.

Page last modified: September 16, 2026