Hi  @MrHC1983 

Thanks for posting here.

“all VPN stops working after ~40–50 days of uptime, and a reboot fixes it”—usually doesn’t mean the VPN feature itself suddenly broke. More commonly, the device enters an abnormal state after long runtimes combined with sustained external traffic.

Temporarily opening the SMTP port and then getting heavily scanned/attacked can significantly increase NAT sessions, connection-tracking (conntrack) table usage, stateful firewall workload, and logging overhead. If the attack traffic creates many short-lived connections or malformed/unusual packets, it can trigger resource exhaustion (memory pressure, conntrack table full, high CPU), hit edge cases in the driver/kernel network stack, or cause a process hang. That can then impact IPSec/L2TP/OpenVPN tunnel negotiation and forwarding, presenting as “all VPN down,” sometimes along with packet loss.

A reboot clears state tables and resets networking processes, so it recovers temporarily.

Typical recommendations are: keep unnecessary ports closed, enable rate limiting/IPS where available, monitor conntrack and memory usage, and stay current on firmware to address long-uptime stability issues.