Troubleshooting

Common issues and solutions for the Exaviz Management Platform

  6 minute read  

Cannot Access the Dashboard

Board IP not responding

  1. Verify the board is powered on (check LEDs)
  2. Confirm your computer is on the same network as the board
  3. Try pinging the board: ping <board-ip>
  4. If using the Cruiser’s DHCP server, check your network’s DHCP lease table for the board’s IP
  5. Try connecting directly via Ethernet cable

Browser shows “connection refused”

The management service may not be running:

ssh user@<board-ip>
sudo systemctl status exaviz-mgmt
sudo journalctl -u exaviz-mgmt --no-pager -n 50

If the service is not running, start it:

sudo systemctl start exaviz-mgmt

Browser shows HTTPS certificate warning

This is expected with the default self-signed certificate. Click “Advanced” and “Proceed” (Chrome) or “Accept the Risk” (Firefox) to continue. See Security and TLS for installing a trusted certificate.

Login Issues

Forgot password

There is no password reset through the dashboard. To reset:

ssh user@<board-ip>
sudo rm /var/lib/exaviz/auth.json
sudo systemctl restart exaviz-mgmt

The password will revert to the default (password) and you will be required to change it on next login.

“Too many login attempts”

Login is rate-limited to 120 attempts per minute per IP address. Wait 60 seconds and try again with the correct password.

“Must change password” loop

Until the default password (password) is changed, the board refuses every change. Enter password as the current password, then choose a new password with at least 8 characters.

PoE Port Issues

Port stuck in “Searching”

The PSE is looking for a PoE device but not finding one:

  • Verify a PoE-capable device is connected
  • Try a different Ethernet cable (Cat5e or Cat6 required)
  • Check that the device supports 802.3af or 802.3at
  • Some devices only negotiate PoE on specific wire pairs. Ensure the cable is wired correctly.

Port in “Fault” state

The PSE detected an error condition:

  • Short circuit - Check the cable for damage
  • Overcurrent - The device may be drawing more power than the port class allows. Try switching to AT mode (30W).
  • Overtemperature - The PSE controller is too hot. Improve ventilation.

Try resetting the port. If the fault persists after multiple resets, the issue is likely in the cable or connected device.

Port cycling between “Backoff” and “Fault”

The PSE is retrying after repeated faults. This usually indicates a hardware issue:

  • Replace the Ethernet cable
  • Try a different PoE device on the same port
  • Try the same device on a different port

If the problem follows the device, the device may have a faulty PoE receiver. If the problem follows the port, contact support.

Port shows 0W despite “Delivering”

The port is powering a device but the power reading has not updated yet. Wait a few seconds for the next SSE update. If the reading remains at 0W, the device may be drawing very little power (below the measurement threshold).

Dashboard Issues

Dashboard shows stale data

The LIVE indicator should pulse green. If it is gray or missing:

  • The SSE connection may have dropped. Refresh the page.
  • Check your network connection to the board.
  • Check if the management service is still running on the board.

Dashboard is slow or unresponsive

  • Check the board’s CPU and memory usage in the health section
  • If another application (NVR, recording software) is consuming resources, the dashboard may be affected
  • Restart the management service: sudo systemctl restart exaviz-mgmt

Fleet Issues

Fleet view does not appear; “N boards” pill but no Fleet tab

Fleet management is opt-in. Confirm /etc/exaviz/management.json has:

"fleet": { "enabled": true, ... }

If enabled: false, mDNS discovery still runs (which is why the “N boards” pill counts peers), but the Fleet view and pair flow are gated until you flip it. Edit the file, then sudo systemctl restart exaviz-mgmt. See Fleet Management - Enabling.

Board not appearing in fleet discovery

  • Both boards must have fleet.enabled: true in management.json (see above).
  • Both boards must be on the same network subnet.
  • mDNS (port 5353 UDP) must not be blocked by a firewall.
  • Some managed switches and enterprise network equipment block multicast by default. Check your network configuration.
  • Wait up to 30 seconds for the announcement cycle.

Discovery is asymmetric (board A sees B, but B does not see A)

If one board is visible to others but does not see them back (or vice versa), the silent board may be sending mDNS announcements out the wrong network interface. This is most common on multi-NIC boards (Cruiser with eth1 + wlan0 + ext0_* + docker0).

Confirm with tcpdump on the silent board:

sudo tcpdump -i any -n 'udp port 5353 and dst 224.0.0.251'

Announcements should leave the LAN-facing interface (eth1 on Cruiser, wan/wlan0 on Interceptor with WAN-side cable, poe0 on Interceptor with WAN cable on a PoE port). If they go out an internal trunk like ext0_2 (link-local 169.254.x.x) or docker0, early builds of fleet management had a bug here that selected the wrong interface. Upgrade to the latest exaviz-management package (sudo apt update && sudo apt install --only-upgrade exaviz-management exaviz-management-kiosk), which announces on every viable LAN interface.

Pairing confirmation codes do not match

If the 6-digit codes shown on the two boards do not match, do not confirm the pairing. This could indicate a man-in-the-middle attack or a network issue. Cancel the pairing and try again.

Paired board shows “Offline”

The fleet aggregator polls each paired board every 5 seconds. If a board is offline:

  • Check that the board is powered on and connected to the network
  • Verify HTTPS is accessible between the two boards (port 443)
  • Check if the management service is running on the remote board
  • A board not seen for 120 seconds is marked offline

Config drift shows on the Fleet view

Expand the drift panel and read the row. HSTS max-age can be fixed from the panel, except that a value above 0 is refused while the board has the HTTP to HTTPS redirect off (turn the redirect on first; 0 is always allowed). For the other settings, follow the steps in the row on that board. If a board is listed under “Unreachable peers:” (the header then says “cannot verify”), check that it is online and has fleet management turned on. After a hand change to /etc/exaviz/management.json, restart the management service, because the check reads the running settings. The check runs every 5 minutes. The drift alert clears on its own at the next check that finds no drift and reaches every paired board. While a paired board does not answer, the alert stays open, because that board may still differ.

TLS Issues

ACME certificate request fails

  • The board must have a public DNS name resolving to its IP
  • Port 80 must be accessible from the internet (HTTP-01 challenge)
  • Check DNS propagation: dig <hostname>
  • Check firewall rules allow inbound HTTP

Browser still shows old certificate after upload

New certificates apply to new connections only. Existing connections continue on the old certificate until they close. Close and reopen your browser, or open a new incognito/private window.

Getting Help

If these steps do not resolve your issue:

Last modified October 18, 2026