Questions[windows-nlb](server)

We are using Windows NLB in our production environment to load balance https requests to ten IIS servers.

The settings are:

So as long as there is no change in the NLB cluster membership, i.e. no hosts are leaving/joining the cluster, the same source ip will always be mapped to the same target IIS server (according to the docs for Server 2003 R2 - I can't find newer official docs that detail affinity):

Now what if I take down some or all machines, and bring all of them up again, so the cluster membership is the same as it was before taking down machines. Will the same source ip addresses again map to the same target servers as before, or does the mapping depend on things like cluster join order or something non-deterministic?

N.B. I just found that during convergence, the mapping may change:

In normal operations, if single affinity is used, NLB will guarantee that all connections coming from the same source IP will hit the same server. This would include multiple TCP connections of the same SSL session. However, configuration changes might cause NLB to accept different connections of the same SSL session by different servers during the convergence. As a result, the SSL session is broken.

Which is fair enough (and can be mitigated by enabling "extended affinity" using the "timeout" value, as the linked article explains). However I'm interested in the state after convergence.

I'm having an application hosted on a VM server2012r2. This application has a webpart, running .NET 5 (Kestrel). The webpart is accessible via http://<serverip>:808 or https://<serverip>:4434.

I have setup a second VM, a server 2019. The application is also installed there (.NET 5, Kestrel), where the webpart is also accessible on the server via port http:808 or https:4434.

The application is configured in such a way that they point to the same database etc...
So basically it are two applications working together but both accessible via their own portal.

The only problem I have is there are two portals pointing to the same:
http://<server2012>:808 and http://:808`

(or the same for https for ports 4434).

The idea is that I enable Microsoft network load balancer so that
(A) the end user only has to go to http://<something>:808 or https://<something>:4434 instead of remembering one of the two portal IPs;
(B) the traffic is load balanced between both (and if one is not accessible, it goes to the one available (i'm not sure if nlb is capable of doing this?)).

I've followed a tutorial as this is the first time i'm using microsoft nlb. https://www.serverlab.ca/tutorials/windows/web-servers/load-balancing-web-servers-with-windows-server-2012-r2/

The problem that I am facing is that when I create a New Cluster with Parameters, IP Address(192.168.0.88 ) and Full internet name (nlbtest.com) and add for the Cluster IP 192.168.0.88 a TCP port from 808 - 808 that I can't access my server anymore via remote desktop.

server 2012r2 (VM 1 NIC): 192.168.0.90
runs kestrel web app on http://192.168.0.90:808 or https://192.168.0.90:4434

server 2019 (VM 1 NIC): 192.168.0.91
runs kestrel web app on http://192.168.0.91:808 or https://192.168.0.91:4434

Cluster IP: 192.168.0.88 for TCP port 808
when I configure this on server2012r2 via microsoft nlb I don't have any connection anymore. I can only use hyper-v manager on the host to access the VM and in the VM I can't ping or surf to the web anymore.

What am I doing obviously wrong?

Currently we have a working NLB cluster of web servers running Windows Server 2008 R2 and IIS 7.5 with one virtual IP address.

We have three web servers in the cluster the VIP is 141, and the web servers each have an IP address of 142, 143 and 144.

We need to add another web site bound to a seperate IP address - we need to do this because we need a seperate and completely different SSL Certificate for the new site and IIS doesn't support two certificates on the same site (ref Thread - Multiple SSL Certificates on a Web Site in IIS 7).

One option for us is to add another VIP say 151, and add new local addresses - for example - 152, 153 and 154, and then make a new site bound to new IP(s).

Please can you tell me, if this is possible and is it a supported production configuration?

I have a questions about load balancing configurations. Maybe someone already asked this question but I didn't find something similar on the web.

The Actual Configuration is .

I Have two servers that I want to load balance.

Server 1 has two NIC

NIC 1 IP :xx.xx.xx.35 Gateway : xx.xx.xx.33
NIC 2 IP 192.168.2.11 

Server 2

NIC 1 IP :xx.xx.yy.45 Gateway : xx.xx.yy.43
NIC 2 IP 192.168.2.10

So server 1 and 2 are connected with the second NIC

I Want to load balance those two server using the NIC 2 But all the traffic coming from NIC1 should be reparted between the two servers.

And until now and didn't find any way to do it . I can't really find information if it's possible or not.

We have a 2 node NLB cluster. I recently added more IP's to the cluster range, after which point, the NLB NIC's on both of my nodes ended up in a permanent "acquiring network address" state.

The cluster appears to be functioning despite this hang in acquiring an address.

We've also rebooted the machines, and these NIC's remain in this state "acquiring network address" even after reboot, which I find troubling.

Questions:

1. Does anyone know where I can find documentation on how NLB binds to a NIC? (are there dll's involved, drivers, etc.)
2. Does anyone know of any programs that have known incompatability issues with NLB?

We have an application server installed on these NLB nodes (AOL Server) that interacts with the NIC's as well, and I suspect that it might be causing this issue. Before I go directly to Microsoft, I want to gather as much information as possible about the issue.

Any help / guidance would be much appreciated...