First things first: Ive been encountering some serious internet problems,with either my gateway or the way the network was set.
I always felt a huge drop in throughput.
After some searching I found out, that netctl can handle custom mtu configurations as well. The 'ExecUpPost' entry in a connection profile handles this. Here's the simple ethernet-static profile that I'm using for my work network :
nerd@trojan:/$ sudo nano /etc/netctl/my_lan
Description='Science Block private ethernet connection'
ExecUpPost='/usr/sbin/ip link set enp5s0 mtu 258'
#And then restart for settings to take effect
nerd@trojan:/$ sudo netctl my_lan restart
PS: before setting the MTU size, initially test and determine the optimum size
optimal MTU value has nothing to do with modem or connected computer. It depends on your ISP
You can check optimal MTU value pretty easy by pinging any website with different MTU value starting from lower going up to first ping error:
start with lets say value 1460 and go up:
In my case 230 has got no error:
nerd@trojan:/$ ping -c 5 -M do -s 230 mmphego.wordpress.com
PING lb.wordpress.com (18.104.22.168) 230(258) bytes of data.
238 bytes from wordpress.com (22.214.171.124): icmp_seq=1 ttl=44 time=1681 ms
238 bytes from wordpress.com (126.96.36.199): icmp_seq=2 ttl=44 time=1961 ms
238 bytes from wordpress.com (188.8.131.52): icmp_seq=3 ttl=44 time=2235 ms
238 bytes from wordpress.com (184.108.40.206): icmp_seq=4 ttl=44 time=1879 ms
238 bytes from wordpress.com (220.127.116.11): icmp_seq=5 ttl=44 time=1793 ms
--- lb.wordpress.com ping statistics ---
5 packets transmitted, 5 received, 0% packet loss, time 7537ms
rtt min/avg/max/mdev = 1681.069/1910.091/2235.116/187.261 ms, pipe 3
and remember: pack_size +byte_size = mtu or something along those lines
in my case: 230+28=258(mtu)
I'm sharing it here, in case
a) anyone else encounters the same problem and
b) I forget again, how to set it