Problem: During the earlier disk-full incident, the agent’s local cache files got truncated to 0 bytes while trying to write to a full disk:

• /usr/local/ltechagent/state
• /usr/local/ltechagent/state_old
• all files under /usr/local/ltechagent/monitors/

The agent kept limping along until the reboot, then failed to start at all because it couldn’t parse an empty state file — and since the service has no auto-restart configured, it just sat failed from then on.

Fix — moved the corrupted files aside (not deleted) and restarted the service:

sudo mkdir -p /usr/local/ltechagent/corrupt-backup-20260813
sudo mv /usr/local/ltechagent/state /usr/local/ltechagent/state_old /usr/local/ltechagent/corrupt-backup-20260813/
sudo mv /usr/local/ltechagent/monitors/* /usr/local/ltechagent/corrupt-backup-20260813/
sudo systemctl restart ltechagent.service

The agent’s actual identity/config (agent_config) was untouched — only the runtime state cache was corrupted — so on restart it rebuilt fresh state using its existing registration and reconnected to pams.pahosted.com.au without needing a reinstall.

Let me know if you still want Restart=on-failure added to the unit file so it self-heals if this happens again.