Cluster Servers?

#1 · 02/25/2024, 5:16 PM

I can see that there is an option to make cluster servers, but I don't know how to do it. Could someone let me know how to do it?

Rumo has reacted to this post.

#2 · 02/25/2024, 8:19 PM

Quote from Administrator on 02/25/2024, 8:19 PM
Hey, please follow the official wiki here

More About Cluster Files and Running Multiple Servers

The cluster files will always be saved under the <ClusterDirOverride>/<clusterid> directory. This implies a few things:

For a specific player, downloading (thus, also, transferring) a Survivor onto a server always overwrites whatever Survivor data was previously there. Inversely, there can only one Survivor 'Uploaded'.

If two servers share the same clusterid but they do not share the same cluster directory (different <ClusterDirOverride>), players will still see on transfer list such servers but they will not be able to download (and thus also transfer) anything uploaded from a server to another.

If you select a clusterid that is common, there is a high chance that your server list eventually becomes "polluted" by other servers that are configured to use the same clusterid which can confuse players since the data is not shared with those other servers. Given this issue, it's suggested to not use a common clusterid name as other unconnected servers could be listed while the players are using the transfers.

Advanced cluster managements clarifications

The following considerations are left for experts administrators only and require proper testing before put such setups in production:

It is possible to set a shared location between different physical or virtual machines to use a common clusterid. If over a netwoek, it might slightly increase the chance of failing transfers given network issues, which typically ends with the player losing the entire data. For such more advanced setup, it would strongly suggested to use a third-party applications to dynamically handle cluster and player data backups.

To spare some drive space and installation/update time, it is possible to use the same installation to run multiple servers. This can also be combined with ?AltSaveDirectoryName.

Note in this way they will also share the same INI settings because they load the same files unless using complicated file linking tricks. To still have them differ it is possible to set most of GameUserSettings.ini options in the command line using the ? syntax. You could also use a dynamically generated INI file to change some of the Game.inisettings. Manual editing of INI files is usually not recommended. The server no longer loads any INI data once the PID of the process is displayed in the output console. It is thus needed to be careful to load the server with proper intervals as not to use the wrong INI options from a previous server launch. Also, it must be remembered that, when quitting the server normally, the process overwrites some content of the INI files with its current settings.

The server deletes old savegame files using the age of the file as long as they are named .ark. This means that regardless of the map you are using, the savegame files will eventually get deleted from the directory. If you stop a map, no more autosaves will be made so leaving any .ark files there will get them eventually deleted. Note that if this happens, you can at least still use your .bak backup files when the server was last started, which do not get deleted, given you don't have another backup solution running on the server.

If clusterid is not set and if -NoTransferFromFiltering is not set: it is still possible for players to use the 'download'/'upload' buttons on terminals. Tribute data would then be stored in players' own machines, the same way as their single player data. This would also allow them to download it into another server with such setup.

If clusterid is not set and if -NoTransferFromFiltering is also set: player transfers from single player is disabled. Given the Tribute Data is saved in the SavedArks directory, it is still possible to create a "fake cluster" by sharing the SavedArks directory between servers, allowing players to use the 'download'/'upload' buttons on terminals. Finally, it's required by players to known what servers are linked.

It is not possible to mix and match these types of servers between one another, as the Tribute Data format will not be the same: with the clusterid, the Tribute Data is saved in clusterid directory and if -NoTransferFromFiltering is not set, it will save on the players' own machines.

Hey, please follow the official wiki here

More About Cluster Files and Running Multiple Servers

The cluster files will always be saved under the <ClusterDirOverride>/<clusterid> directory. This implies a few things:

For a specific player, downloading (thus, also, transferring) a Survivor onto a server always overwrites whatever Survivor data was previously there. Inversely, there can only one Survivor 'Uploaded'.
If two servers share the same clusterid but they do not share the same cluster directory (different <ClusterDirOverride>), players will still see on transfer list such servers but they will not be able to download (and thus also transfer) anything uploaded from a server to another.
If you select a clusterid that is common, there is a high chance that your server list eventually becomes "polluted" by other servers that are configured to use the same clusterid which can confuse players since the data is not shared with those other servers. Given this issue, it's suggested to not use a common clusterid name as other unconnected servers could be listed while the players are using the transfers.

Advanced cluster managements clarifications

The following considerations are left for experts administrators only and require proper testing before put such setups in production:

It is possible to set a shared location between different physical or virtual machines to use a common clusterid. If over a netwoek, it might slightly increase the chance of failing transfers given network issues, which typically ends with the player losing the entire data. For such more advanced setup, it would strongly suggested to use a third-party applications to dynamically handle cluster and player data backups.
To spare some drive space and installation/update time, it is possible to use the same installation to run multiple servers. This can also be combined with ?AltSaveDirectoryName.
- Note in this way they will also share the same INI settings because they load the same files unless using complicated file linking tricks. To still have them differ it is possible to set most of GameUserSettings.ini options in the command line using the ? syntax. You could also use a dynamically generated INI file to change some of the Game.inisettings. Manual editing of INI files is usually not recommended. The server no longer loads any INI data once the PID of the process is displayed in the output console. It is thus needed to be careful to load the server with proper intervals as not to use the wrong INI options from a previous server launch. Also, it must be remembered that, when quitting the server normally, the process overwrites some content of the INI files with its current settings.
- The server deletes old savegame files using the age of the file as long as they are named .ark. This means that regardless of the map you are using, the savegame files will eventually get deleted from the directory. If you stop a map, no more autosaves will be made so leaving any .ark files there will get them eventually deleted. Note that if this happens, you can at least still use your .bak backup files when the server was last started, which do not get deleted, given you don't have another backup solution running on the server.
If clusterid is not set and if -NoTransferFromFiltering is not set: it is still possible for players to use the 'download'/'upload' buttons on terminals. Tribute data would then be stored in players' own machines, the same way as their single player data. This would also allow them to download it into another server with such setup.
If clusterid is not set and if -NoTransferFromFiltering is also set: player transfers from single player is disabled. Given the Tribute Data is saved in the SavedArks directory, it is still possible to create a "fake cluster" by sharing the SavedArks directory between servers, allowing players to use the 'download'/'upload' buttons on terminals. Finally, it's required by players to known what servers are linked.
It is not possible to mix and match these types of servers between one another, as the Tribute Data format will not be the same: with the clusterid, the Tribute Data is saved in clusterid directory and if -NoTransferFromFiltering is not set, it will save on the players' own machines.

Rumo and WM Plays have reacted to this post.

#3 · 02/29/2024, 4:11 PM

K thx for letting me know!

Administrator has reacted to this post.

#4 · 04/02/2024, 4:10 PM

Unfortunately, the text in English is too complex for me. And the meaning is lost through translation programs.

I am now reading through the instructions for clusters and hope to understand something. The ASADecicatedManager has unfortunately become very confusing and I no longer understand the connections.

I get the message that the RCON port is causing a conflict and each server needs a port. I just don't know what exactly that means? Do I need to open a 27021 and 27022 on 3 servers in addition to 27020? How do I know that these are not occupied by anything else? Can't the software recommend which PORTS are not yet in use and release them with a click?

What is "MultiHome" and why can it be activated or deactivated? Only my IP is specified there and I cannot select another one. Unfortunately not intuitive for noobs.

Can I simply move existing server paths to the cluster path?
Where is the point where I can reselect the server path? It was displayed at the top at noon at the beginning. It is still displayed, but without the option to change it!

Translated with DeepL.com (free version)

#5 · 04/05/2024, 11:35 AM

it is up to you to configure the rcon port you want in the GameUserSettings tab rcon section each server (profile) must have a different port because the manager communicates with the dedicated server via rcon to stop it or save the world you do not need to redirect this port on the internet in your router and I do not recommend opening this port on the internet it is used locally by the manager unless you want to use rcon from outside your local network

The Multihome command is used to specify the IP address that the server should use to communicate externally. This is particularly useful in environments where the server is hosted on a machine with multiple network interfaces or multiple IP addresses assigned. Using the Multihome command, you can force the server to bind incoming and outgoing network traffic to a specific IP address. This allows for better bandwidth management and more precise network configuration, especially if you want to separate gaming traffic from other types of traffic on your server or if you need ARK's server to be accessible over an IP specific public address in a NAT (Network Address Translation) environment.

do not use it if you do not know that you do or do not have several network interfaces

Rumo has reacted to this post.

#6 · 04/06/2024, 7:52 AM

Many thanks for the answer.
I think I got it working with rcon and the ports. Also, all 3 servers start in the cluster, but I always get a timeout in the game itself. Likewise my son from another network. I'm a bit desperate because I don't know what the problem might be.
My cluster with the maps is located at C://clusterfolder/
Not under the directory where the ASAManager is installed. Is this a requirement?