Anyone using Windows Network Load Balancing with multiple connection servers? Utilizing Win 2008 r2 standard
yes we did it in 4.6 and in 5.0. Works just fine for us.
Yes, I was doing this with 4.6, we are still doing it in our 5.0 environment.
Any issues with the hosts not recognizing when one became unresponsive?
I'm not sure what you mean? I haven't had anything become unresponsive, so I guess my answer would be no, no issues for me. I've been running this since February. I'll be changing for a clustered pair of load balancers in the future though as traffic increases.
Hi csterley,
Where using a MS NLB cluster installed on W2K8 R2 in unicast mode. We encounterd no issues so far on View 4.6 and since last week View 5.0.
Only issues we had was solved by following the instruction in KB: http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1556. We tested it extensively.
NLB does not monitor the underlying VMware infrastructure including View Manger services ect. So, if the server is replying to ICMP requests and the Windows side is "all good" but the View side is having trouble, clients will still be sent there only to fail at connecting to the manager.
To get full load balancing and redudancy you'll need to look at a more sophisticated load balancing solution that can actually test a login to the connection server among other things.
FYI - We are also using MS NLB to load balance but we monitor the other pieces of the infrastructure with another product (Solarwinds NPM). We have not specifically had any issues with the NLB side but if any of the required View components on the server fail or are not running, clients are still sent there and attempt to connect.
Hope this helps.