Why bind cluster and client/server to different IPs?

Binding to different IPs is a best practice for network isolation. It ensures that intensive internal cluster data replication (state transfer) does not compete for bandwidth with external client-to-server requests, significantly reducing latency and improving overall cache throughput.

How can I configure NCache to use specific IPs for different types of traffic?

You can configure NCache to use specific IP addresses for cluster and client/server traffic through the NCache Management Center. By selecting the desired cache node, you can assign a dedicated IP for cluster communication or client/server communication. This ensures that NCache uses the correct network interface for each type of traffic.

Does changing NIC binding require a cache restart?

Yes, changing the IP binding for an active node requires a restart of the cache instance on that specific node to initialize the NCache Service on the new network interface.

What should I do if NCache nodes cannot communicate after binding to different IPs?

If NCache nodes cannot communicate after binding to different IPs, it may be due to mismatched IP configurations. To resolve this, ensure that all nodes are bound to publicly accessible IPs that can be reached across the cluster. You can also adjust the NCache service configuration file to bind the service to an IP that is reachable by all nodes, helping to avoid communication failures.

Bind IP with Multiple NICs

When using NCache on servers with multiple Network Interface Cards (NICs), you can configure the binding of cluster communication (node-to-node) and/or client/server traffic to specific IP addresses. This allows you to separate internal cluster traffic from client-facing communication, improving security, performance, and network isolation. Binding to dedicated IPs ensures that private cluster traffic remains isolated from public client traffic, which is particularly useful in environments with separate public and private networks. This configuration can be done through the NCache Management Center or directly in the service configuration file, depending on your preference and setup.

Warning

Binding cluster and client with different dedicated IP addresses is only useful if the server machine has multiple NICs configured.

A cluster or a specific client/server can employ dedicated IP binding using the NCache Management Center as detailed below:

Using the NCache Management Center

Launch the NCache Management Center by browsing to http://localhost:8251 or <server-ip>:8251.
In the left navigation bar, click on Clustered Caches or Local Caches, based on the cache to configure.
Select the cache name, click on View Details.

NCache Management Center showing the 'View Details' button for selecting a cache configuration.

This opens up the detailed configuration page for the cache.
Under the Server Nodes tab, check the box against the server node IP to change NIC.
You can do so by using either of the following methods:
- Method 1: Click on the button on the toolbar to change NIC.
- Method 2: Click on "..." against the server IP and select Change NIC.

Detailed configuration page for a cache node in NCache Management Center showing the option to change the network interface card (NIC).

Server node started web

Upon either selection, the Configure NIC page will open. It will list all network cards available on that node with their assigned IP addresses.
Select the Cluster or Client/Server option as per your requirement, from drop-down in front of the IP with which you want.

Configuration page in NCache Management Center for binding cluster or client/server traffic to a specific NIC with a dropdown selection.

On clicking OK, you will receive the following prompt warning you about the running caches.

Warning prompt in NCache Management Center confirming the NIC change for cluster or client/server traffic binding.

Click OK to complete the binding process.

Troubleshooting Dedicated IP Binding Clusters

If you have dedicated network cards configured for public and private communication, then you might not be able to create new clusters. Additionally, it is possible that existing NCache clusters could fail to communicate. Please note, that the following problems will occur if nodes are unable to communicate with each other on specific interface(s).

The NCache Service starts at one interface, whereas other nodes try to communicate with it on the other interface. Therefore, the two nodes are always unable to communicate with each other even when both nodes are fully functional.
The NCache Clients connect to the NCache Server to get OutProc cache instances. So, the NCache Server starts at one interface, whereas the client tries to connect to it on the other interface. This way, though the NCache Server is ready to accept the client requests, the client applications will always be unable to connect to the OutProc caches.
Situations can arise where the cluster nodes are unable to join with each other and no cluster forms. This could stem from one node listening to join requests from other nodes on one network interface, while the other nodes send these join requests on the other network interface.

Workaround

To avoid such problems, you can bind the NCache Service to a specific IP that is publicly available to other nodes in the cluster. You can do this in the service configuration file in %NCHOME%\bin\service or opt/ncache/bin/service (for Linux). Please note that for .NET Core Linux installation, the file name is Alachisoft.NCache.Daemon.dll.config.

<add key = "NCacheServer.BindToClientServerIP" value = "<your-ip>" />

Bind IP with Multiple NICs

Warning

Using the NCache Management Center

Troubleshooting Dedicated IP Binding Clusters

Workaround

See Also

Contact Us