CAS Array Failover Issue

I am owner of two sites " A & B". Site A will be the live site and established two CAS/HUB servers configured within a CAS Array and 2 MBX servers. Where as Site B established with a single CAS/HUB server as well as a single MBX server. Entire three MBX servers are configured within a DAG array and the CAS/HUB in Site B will only be implemented in case of a complete Site A failure. Last night, I have checked the efficiency of the CAS array by turning off one of the CAS/HUB servers. When the server turned off, we were not able to open up the OWA or do any client access functions. As soon as, it came back up all of thing and was well.

I executed the get-mailboxdatabase | fl *rpcclientaccess*,name and all my databases sections to the CAS Array. What can I see at as a cause for my NLB CAS array getting failed when turning off just one of the servers within the array.

Would you provide me some information regarding the same means the workaround and all . Does the DNS A Record for your CAS Array indicate to the VIP of the Load balancer specified by you ? If you have suddenly indicated this then the host address of the CAS Server , you turned off then you will get this problem.

One more thing, I need to know here ...means what Load Balancing fix are you working with ? As this could be a configuration problem with your priority, affinity or service health testing which is still trying to send request onto the failed host.

Thanks for your suggestion and I am so glad that you viewed my question and spend your valuable time to provide me the solution . Actually, the CAS array DNS A record does indicate to the VIP of the specified load balancer. We are currently working with the windows load balancing fix. I did got within my Advanced TCP/IP configurations on the NLB NIC's and I am running with two different IP's. One obviously is the NIC enabled IP Address , but the second one is specified as VIP. Do I require the VIP IP combined with the NLB NIC's ?

Hi guys !! I am running with two CAS/HT servers as the virtual machine in a CAS array. The array displays healthy and well . When I turn off the primary CAS, pings stay live and the another displays as the master.

However, when the primary CAS server got tun down , Outlook clients found a send/receive error 0x80190193 but I tend to consider that is it occurred just because I have not transitioned upon the OAB stuff yet (still checking). However, what I am not sure about this and need to be aware about this what should happen in a failover? Is this assumes to be seamless,all of thing is going well as if nothing happened? There is No logouts there and not any type of errors, etc ?

The primarily established CAS server is a physical server where as the second is one of the virtual server within a vSphere environment. The server, we turned off was the virtual server. I have just scanned the msexchange.org:

"When you decided to set the WNLB array for virtual Exchange 2010 CAS servers which is based on the VMware ESX Server to perform the required task as the virtualization platform, it is suggested to set your WNLB in multi-cast mode since you otherwise will expect a problem with the WNLB array not functioning in well manner "

I have tested the Cluster Operation Mode and we have configured with the Unicast. What is your idea about this on the Unicast configuration causing it ?

Did you got an answer for this, I am running with the mixed performance and connectivity problems with the same configuration (WNLB and Unicast for two CAS servers working on to the ESX 3.5 Update 5) I haven't accomplshed the steps for the VMWare KB (http://kb.vmware.com/selfservice/mic...xternalId=1556) becuase this is the point which effects the whole VSwitch that my servers are working on. I need to just switch my WNLB to Multicast as a easier fix in case, if its also a suggested solution for my occurred issue.

What is the actual definition on NLB NIC ? As I am sure , in the unicast mode, each node contains at least two NICs, one (Public NIC) contains the IP subnet that all of the servers/clients are able to communicate to; and the different (NLB NIC) contains the IP subnet which only NLB nodes are able to use to communicate with each other

If your definition is the similar, the VIP should configure on public NIC, not onto the NLB NIC. The NLB NICs are only implementing for the NLB communication among two nodes cross-over or through the VLAN if you will associate more node to NLB cluster in the future in some cases.