So, I got the server setup, file server in place, media server in place, rsync backups taken care of, a lot of network configuration on the router for a bunch of this shit and thought I nixed my packet drop issues.
But here they are, still here.
So if anyone remembers the blog when I showed this:
It is now alive and instead of running a VM for the game server I decided why not just use this box for it instead? Maybe get a VPN setup for emulation multiplayer without having to dedicate some time, how cool would that shit be? This will be the machine used for all my game servers. (I’m going to a lan in June, so hopefully I can knock out any kinks by then.)
And while the other server has some other alarms going (A friend of mine told me about netdata, which is where I’m getting all this info from) none of those are actual problems just telling me that my servers are 5-6 year old hardware
If I run
netstat -s | grep retransmitted
I would initially get something like 1000 segments retransmitted after an uptime of an hour. I unplugged one of the ethernet cables and rebooted. I thought maybe there’s more configuration I needed or having two IP addresses is screwing it up, but still dropping packets.
Here’s the segments retransmitted for my desktop though so not sure there’s really that much of an issue.
1847 segments retransmitted 1299 fast retransmits
I checked my routes and it goes like
Kernel IP routing table Destination Gateway Genmask Flags Metric Ref Use Iface default gatekeeper.leon 0.0.0.0 UG 100 0 0 enp3s0 default gatekeeper.leon 0.0.0.0 UG 100 0 0 enp2s0 192.168.1.0 0.0.0.0 255.255.255.0 U 0 0 0 enp3s0 192.168.1.0 0.0.0.0 255.255.255.0 U 0 0 0 enp2s0 gatekeeper.leon 0.0.0.0 255.255.255.255 UH 100 0 0 enp3s0 gatekeeper.leon 0.0.0.0 255.255.255.255 UH 100 0 0 enp2s0
And I notice something weird, that my domain destination is on /32 instead of the actual subnet the computers are on /24 (I’m a basic ho I know) and that the gateway is empty. I have an internet connection. I can SSH into everything. Netdata shows in browsers and plex is still running on the other box. I’m not sure if this really is weird or not. It’s kinda new to me.
So I’ve gone into pfsense and changed the gateway ip of the DHCP server to 192.168.1.1. And that seems to have fixed it? I’m still getting a warning, but There’s no actual change to the gateway in the routing table. The download from steamcmd is going much faster than it used to, the last time I tried it failed, and the time before that it took an actual hour, this time it completed in about 15 minutes. Both cables are in and link is up and after about an hour I am down to
27 segments retransmitted 1 fast retransmits
The server itself is ubuntu server 18.04, running some 5000 series xeons on a TYAN S7002 motherboard with whatever revision is the really stripped down model. I have turned off the software firewall in ubuntu for testing purposes. Thinking about maybe switching to centos or fedora. But if I can actually just keep doing I’ll just run a VM for getting familiar in the red hat distros.