Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Cluster on two different servers(different Pc's) #993

Closed
Stray-Fenris opened this issue Dec 14, 2018 · 8 comments
Closed

Cluster on two different servers(different Pc's) #993

Stray-Fenris opened this issue Dec 14, 2018 · 8 comments
Labels
Support Issues that aren't bugs

Comments

@Stray-Fenris
Copy link

I have switched to a newer Pc(was my gamepc,then a few years as a company-pc and got now replaced, so back in my hands), which has 8 cores, so i wanted to run two servers, sadly, i realised now, that it has only 8GB Ram max, which is not enough to start two non-Dlc servers, not even talking about 2 servers with EC.

So here i am, debating wether i should reactivate my old server as well or if there is another solution. Also, i would love to know how the cluster function works, because it's most likely, that i will need to use it

@klightspeed
Copy link
Collaborator

The instances need access to a shared directory, which contains data passed between the instances in the cluster.

@Stray-Fenris
Copy link
Author

Stray-Fenris commented Dec 14, 2018

@klightspeed Sooo, would a external harddrive in the network do the trick?Edit: I should mention, they wouldn't be running non -stop but on-demand, for a crew of max. seven people Edit2: and how do you specify, where the directory lies? Or is it the arkserverrot?

@TruWrecks
Copy link

You can map a shared drive for the cluster on a NAS or another PC.

Example: Create a share on 1 PC and use fstab to mount the remote shared directory with read/write permissions. Then you point the alternate save directory to the shared paths on both servers. Make sure they use the same ClusterID.

The downside to doing it this way is it can stop one server from booting if the remote drive is not up and listening when the mount command is executed.

@CaspyBXL
Copy link

Dear @klightspeed

I have an update for you what might lead you on to something or help even others.

I was working on some kind of test.
Its my understanding if the 2 server have the same folder setup and same users and permissions the info does go from one to another cluster. Don't ask me how or what .. it just does ...
I have cluster on SERVER A linked to cluster on SERVER B wich is an exact replica of server A (but not a clone). Just set up and scripted the same way.

My people have been transfering from Server A to B since 3 weeks now.
The servers are Phisical Servers and on completely isolated networks.
I use to do this on VM's in the pas with a shared folder for the servers, but now that it works we can once again enjoy the full power of non virtualised servers with still a way to link clusters.

@Stray-Fenris
I would suggest to specify it in the XXX.cfg inside /etc/arkmanager/instances as demanded

arkopt_ClusterDirOverride="/home/steam/5-way-cluster" --> change this to the /media/XXX for a mounted media or if you use NFS or CIFS you could bind it on your own /XXX at the root of the server.
arkopt_clusterid=XXXXXXXXX

@klightspeed
Copy link
Collaborator

Its my understanding if the 2 server have the same folder setup and same users and permissions the info does go from one to another cluster. Don't ask me how or what .. it just does ...

This suggests that inter-server communication is set up when transferring. I wonder if this means that the cluster ID now needs to be universally unique, and not just unique to the server it's running on?

@CaspyBXL
Copy link

I you are up for some testing I can manage some 2 Linux Servers to test this Klightspeed, i'll be happy to give you some resources if needed. Always happy to help out !

@dahlskebank
Copy link

This suggests that inter-server communication is set up when transferring. I wonder if this means that the cluster ID now needs to be universally unique, and not just unique to the server it's running on?

Cluster ID needs to be universally unique. If you try with like 123, you'll get tons of other servers showing up when you wanna transfer, but due to not sharing write privileges with the cluster-files, your char and items can't be moved there (without enabled filtering setting).

Someone solved the "on different networks" cluster thingy, I assume the same could be applied even if on same network (never tested myself): https://survivetheark.com/index.php?/forums/topic/300528-set-up-server-cluster-on-different-networks/

@NobodyDz
Copy link

NobodyDz commented Sep 14, 2019

Use syncthing https://syncthing.net/
It works great, just a bit tricky setting it up or at least was for me
if need help contact me discord: Dz#3600

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Support Issues that aren't bugs
Projects
None yet
Development

No branches or pull requests

7 participants