Quantcast
Channel: High Availability (Clustering) forum
Viewing all articles
Browse latest Browse all 5654

[2012 R2 NLB] http clients see 1 timeout when NLB node is stopped

$
0
0

Windows Server 2012 R2 Network Load Balancing

Test Environment Configuration:

  • 1 Windows Server 2016 Hyper-V host
  • 2 Windows Server 2012 R2 VM's acting as the 2 NLB cluster nodes
  • each node has a management NIC and a dedicated NLB NIC
  • NLB cluster operation is Unicast
  • NLB cluster IPv4 address has a static A record in DNS
  • each VM config has MAC spoofing enabled for the NLB NIC; I presume if I statically assigned the NLB Cluster's MAC to the NLB vNIC in each VM's config it would work without spoofing enabled - right?
  • inside each VM, the NLB adapter is configured to avoid registering with DNS and has no gateway or DNS servers defined
  • Sole Port rule is port 80 only, tcp/udp, no affinity - I'm simply serving up a file

When I http://nlb.contoso.com/somefile, I get to one node and subsequent page refreshes hit the same node. If I stop that node in NLB Manager, then refresh the page on the client, the page refresh times out. On the next refresh, the page is loaded from the remaining NLB node. Why? I had the impression that as soon as I stop node1, refreshing the page on the client should immediately get served up from the still-running node2.


born to learn!


Viewing all articles
Browse latest Browse all 5654

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>