CAS Array - What is supposed to happen in failover? (Full Version)

All Forums >> [Microsoft Exchange 2010] >> High Availability



Message


elluke -> CAS Array - What is supposed to happen in failover? (21.Jul.2010 9:13:13 PM)

I have two CAS/HT servers as virtual machines in a CAS array. The array appears healthy and fine. When I shut down the primary CAS, pings stay live and the secondary shows as the master.

However, when the primary CAS server is down, Outlook clients get a send/receive error 0x80190193 but I tend to think that is because I haven't transitioned over the OAB stuff yet (still testing).
However, what I don't know and want to understand is what should happen in a failover? Is it supposed to be seamless, everything work as if nothing happened? No logouts? No errors, etc?

I need to know this because I also am not sure if NLB works on now my the vmware ESX architecture. I found numerous issues, forcing me to use multicast and make a bunch of undesired changes. The thing that is really concerning me is that if both CAS servers are up and I ping the Array address, I get two icmp responses, presumably one from each server. I only get this when pinging from a unix box. Windows does not show dup ping.




melry88 -> RE: CAS Array - What is supposed to happen in failover? (22.Sep.2010 8:45:36 AM)

Hello,

I have two CAS servers running Windows 2008 R2 with NLB on XenServer 5.6.

Functionality works fine if I disable the NIC on one of the NLB members the clients stay connected. Email will continue to be queued in the Outbox and once everything has successfully cutover it will send with no interaction from the user.

I have tested this setup failing back and forth with this two node setup.

*****Thanks to Jaap Wesselius's article on this site about AutoDiscovery I found this KB Article that discribes on Hyper-V which might also apply to ESX about how the Hypervisor handles the NLB Cluster Virutal MAC*****


http://support.microsoft.com/kb/953828/en-us

Regards,

Ryan




Page: [1]