Network load balance Windows 2008 R2 for View Manager

[chriswu11111]

Hi All,

I have created a NLB on both view manager servers.

Host 1: 192.168.0.10

host 2: 192.168.0.20

Cluster ip address is 192.168.0.100 i have tried both unicast and multicast mode

I can ping the cluster fine, i can use the cluster address to access the view admin console http://192.168.0.100/admin

I can use view clients to connect to the cluster address and get desktop machines.

But something goes wrong as i'm trying to test the fail over.

If i take down either hosts and leave the other one powered on.

I am then only able to connect for a few seconds to the cluster address then after a while the cluster cannot be accessed anymore ....i can ping the cluster address

but if i try to run the admin console from the cluster address or even try to access the cluster address for the view clients.

It is unable to connect.

Is there any reason for this? am i doing something wrong? Is anyone implementing this?

[caryers]

Hey Chris,

I am currently implementing NLB for my two View 4.6 Windows 2008 R2 64-bit servers successfully. When I pull a server out of the rotation, I perform a DrainStop from the NLB GUI. DrainStop is supposed to migrate the connected users on 1 cluster to the other cluster.

I am trying to implmeent NLB for my two Security Servers, but have yet to verify View is functional on SS cluster node2 successfully.

Are you using NLB for your Security Servers (SS) ...granted if you are implementing remote access?

Thanks,

Scott

[chriswu11111]

Hi No i'm running it on connection servers.

Security servers we won't need, if you have juniper vpn appliance you can actually use this as the security server it is built to support citrix and vmware as a security gateway and will authenticate with the connection server/AD and then point them to a list of desktops that they have been entitled.

I'm having trouble with the NLB.

I thought it should automatically shift over to the other server in the cluster should the original server be down.....

[chriswu11111]

Hi does no one have an answer please?

I have tried runnign unicast and multicast mode. I have tried using 2 nics per server node one for cluster one for its own host address.

I have set the settingson the vmware vswitch side so that it is not notifying the switch or rarp requests.

I am seeing the problem that after some time that the systems are up....

I see several problems of:

- i can ping the cluster address but when i try to access antyhing on that addres e.g. http://clusteraddress/admin for the view manager page it doesnt load.

even thought i can ping the cluster address, secondly view clients cant connect to it.

or the other happens:

I can't ping the cluster address at all........so that means i cant connect to admin page or connect via the view client.....

_________________

In addition by testing the NLB i try to take down one node at a time to see if the cluster works but no it doesnt either one of the above happens......

OR one of the above happens even after a while when both NODES are up.

I really need some help with this can anyone provide any at all please??? there must be some expert or someone who has already got this setup...any pointers would be great for me to use to test!!!

[milton123]

I found this thread useful, Have a look..

http://communities.vmware.com/thread/326402

Cheers, Yours Udin