Weird latency pinging via OpenVPN

[bingo600]

Had a strange experience today on ly L2L connection to one of the remote sites. Both 2.4.5-p1

The ICMP time to sw-01 (via vpn) on the remote site is usually around 220ms , but this morning it exploded to 1000+ ms even 1800ms.

I asked the remote site to check the Inet connection , and they had no alarms.

I dug a bit around in the OVPN logs and saw nothing.

When i got back to the problem after lunch , i still had those issues 100+ ms ping time , and alarms on my Zabbix.

I decided to start allow icmp from my central site to the remote site's WAN ip , and was quite "stunned when it was around 220ms" , but a ping inside the VPN still 1800ms .....

In desparation i restarted the central openvpn daemon from "Services" , and afer a client reconnect. I also had 220ms pingtime inside the tunnel.

I have this in both ends:

fragment 1450;
mssfix 1450;

And have 5 other L2L OVPN tunnels wo. this issue.

What could have happened here , that a "OVPN server restart fixed" ??

Edit: CPU load in both ends approx 5..10% , 20% Ram used

/Bingo