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


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 2 nlb cluster nodes
  • each node has management nic , dedicated nlb nic
  • nlb cluster operation unicast
  • nlb cluster ipv4 address has static record in dns
  • each vm config has mac spoofing enabled nlb nic; presume if statically assigned nlb cluster's mac nlb vnic in each vm's config work without spoofing enabled - right?
  • inside each vm, nlb adapter configured avoid registering dns , has no gateway or dns servers defined
  • sole port rule port 80 only, tcp/udp, no affinity - i'm serving file

when http://nlb.contoso.com/somefile, 1 node , subsequent page refreshes hit same node. if stop node in nlb manager, refresh page on client, page refresh times out. on next refresh, page loaded remaining nlb node. why? had impression stop node1, refreshing page on client should served still-running node2.


born learn!

hi ajm,

after nlb configured, when access multiple times, requests load balanced?

>>if stop node in nlb manager, refresh page on client, page refresh times out.

when close page , open new one, successful load?

best regards,

leo


please remember mark replies answers if help.
if have feedback technet subscriber support, contact tnmff@microsoft.com.



Windows Server  >  High Availability (Clustering)



Comments

Popular posts from this blog

Error: 0x80073701 when trying to add Print Services Role in Windows 2012 Standard

Disconnecting from a Windows Server 2012 R2 file sharing session on a Windows 7,8,10 machine

Windows 2016 RDS event 1306 Connection Broker Client failed to redirect the user... Error: NULL