I created a VPN service using OpenVPN and I was trying to figure out where the bottleneck is because it's very slow compare to the internet speed I have on the server. But during the process I found things very confusing (the sevice is in LXC):
1. speedtest in the PVE host and LXCs have much higher Pings than in VM's (~7mms v.s. ~4mms)
2. speedtest extremly slow in host and LXCs compare to in VM's, ( 6-7 Mbps v.s. ~940Mbps )
3. the VPN service is in one of the LXCs, and when I connect to it using my phone, it's able to hit up to 300Mbps. (This is the one that confuses me the most, my connection to this VPN service within LXC is able to hit 300Mbps, which from my understanding, it means the bottleneck of my phone <=> the LXC <=> VPN processing <=> the testing server is 300Mbps, but I cannot hit even 10Mbps if I run it directly in the LXC) . Besides, if I put the VPN service in the VM, it performs the same (300Mbps).
VM:
the PVE host:
LXC:
(I've tested both vmbr0 and vmbr1 for VM, host and the lxc, they have the same results)
1. speedtest in the PVE host and LXCs have much higher Pings than in VM's (~7mms v.s. ~4mms)
2. speedtest extremly slow in host and LXCs compare to in VM's, ( 6-7 Mbps v.s. ~940Mbps )
3. the VPN service is in one of the LXCs, and when I connect to it using my phone, it's able to hit up to 300Mbps. (This is the one that confuses me the most, my connection to this VPN service within LXC is able to hit 300Mbps, which from my understanding, it means the bottleneck of my phone <=> the LXC <=> VPN processing <=> the testing server is 300Mbps, but I cannot hit even 10Mbps if I run it directly in the LXC) . Besides, if I put the VPN service in the VM, it performs the same (300Mbps).
VM:
the PVE host:
LXC:
(I've tested both vmbr0 and vmbr1 for VM, host and the lxc, they have the same results)
