Traditionally cas failover is done with load balancers but i am wondering if this would work as a way to avoid the cost of purchasing load balancers. We have 2 multi-role exchange 2010 servers one at our main site and one at a DR site. I realize that we can repoint the IP address in DNS to the disaster site but i am trying to avoid the manual intervention and DNS chache issues associated with that. We do not use outlook anywhere at all currently. Could i setup a casarray that points to our main server and point our clients to that, then Setup outlook anywhere with a name of say outlookanywhere.company.com and create an internal DNS record that points to our DR server. i would then setup outlook anywhere in outlook (2007 and 2010) point to outlookanywhere.company.com with the "On fast networks, connect using HTTP first, then connect using TCP\IP" unchecked. If i understand this right when our main server is available it will connect over regular tcp\ip. if the client is unable to connect because the server is down it will connect to outlook anywhere which points to our DR server. Is there any reason this would not work? What would the client experience be like if outlook were open durning a failure would it need to be closed and re-opened or would it failover automatically?
- 已更改类型 wendy_liuMicrosoft Contingent Staff, Moderator 2012年7月6日 6:22
That wouldn't be a CAS failover as such because the protocol is changing from RPC to HTTPS. Also, I doubt that config is supported.
You can try this config, will give you HA but not LB. http://social.technet.microsoft.com/Forums/en-US/exchangesvravailabilityandisasterrecovery/thread/ed5582a0-9ff9-4992-ba48-edef7e7257cd