Customer report
MT Pelagia
“No internet on the bridge or the crew network. Websites won't load at all, but the modem lights look normal.”
Diagnosis
NCM config push rev=1857 changed the DNS servers in config/network.json from 10.255.0.53 to 10.255.0.35 (a transposed typo). Every DNS query times out, so nothing loads although the RF link is healthy. Restore the last-known-good DNS.
Evidence
- Config check: dns primary/secondary 10.255.0.53 -> 10.255.0.35
- messages 03:30:00: config bundle applied rev=1857 source=ncm-push files=config/network.json
- messages 03:30:18 onward: fwd dropped from 3.68 Mbps to 0.02 Mbps with wan=up lan0=up
- messages: repeated 'dns: query timeout server=10.255.0.35' every ~3 min
- Link healthy: rx_lock=1 esno 9.5-10.0 dB (24h avg 9.8), point_err 0.06 deg, BUC 1.78 A / 38 C, GPS 3D fix
- Other B3 terminals normal (2-4.7 Mbps); qos.json and terminal.json identical to LKG
Evidence gaps
- Terminal supports no remote command diagnostics (test dns, ping hub, show interfaces), so DNS failure was confirmed only from logs and config diff.
- Syslog timezone is +01:00 and the log runs slightly ahead of collection time. Event order was consistent across entries, so this does not change the conclusion.
Next steps
- Push the last-known-good network.json (DNS 10.255.0.53) to T-1028 and roll back rev=1857
- Check that DNS queries succeed and fwd throughput returns to ~3-4 Mbps
- Review NCM push rev=1857 to confirm no other terminals received the wrong DNS address
- Ask the NCM team to add validation for DNS changes
Investigation
The investigation trail is sealed. Design partners get the full walkthrough.
Request a walkthrough →