RedisBungee Support

[Scaleblade]

Hey,

So I am here today to try and get some feedback and opinions with an ongoing RedisBungee issue I have been having.

First of all I am going to explain my setup and why we use this.

We currently run 2 Dedicated Machines (Alpha and Beta). These two machines are separated into two different types of servers running on them.

Alpha - 1.10.2-1.12.2 Modded and Vanilla Servers
Beta - 1.7.10 Modded Servers

We have to have a setup this way due to the packet limits put in place by regular Bungeecord for 1.7.10 modpacks resulting in us having to use a customised version of Bungeecord from Waterfall.

Both of these servers are linked together using a Redis database. Now an issue that we have been running into is that the Alpha server has been disconnecting from the redis and cutting the playercount down. When looking through BETA logs you can see this in the history:

20:19:01 [SEVERE] [RedisBungee] proxy2 is 240 seconds behind! (Time not synchronized or server down?)

Proxy2 being Alpha's proxy. Now there is a simple fix of just restarting this proxy however when you have 120+ players on the network it isn't something you want to do - but at the same time I am forced to do it so we can keep our statistics up on vote websites.

Im stumped on what to do? Is it something that could be fixed with some sort of connection checker on a custom forked version of RedisBungee or is this something that I have to deal with? Its not something I see and hear a lot of other servers having though :/

Thanks for any input!
Jack C.

 

[Ajdin]

What does your Redis run on?

What's the load like on all involved machines?

Might sound stupid but perhaps your server times aren't correct?
Try synchronizing your OS server time with the same time server as your other node(s).

 

[Scaleblade]

Both of the servers are hosted in the same data center and run the same OS and times as far as I know of.

The redis for the player count is actually hosted on its own entirely dedicated VPS server also in the same data centre. Load for both servers look very similar if not the Alpha node receives less players than the Beta one right now.

 

[ixilon]

> run the same OS and times as far as I know

That's what you think it is.

> proxy2 is 240 seconds behind! (Time not synchronized

... and that's reality!

 

[sergivb01]

Have you "killed" a bungee server or has any bungee server crashed? Anyways, you should try and remove the redisbungee data from redis with both bungeecords closed and start them again.