DoveSupport
asked on
Annoying Outlook Anywhere
Hi All
Having recently published our Exchange 2010 SP2 CAS Array to host Outlook Anywhere, via NTLM, I'm about to throw them across the room !
We have this scenario :-
OWA :
Public IP1 - Cisco FW - External IP1 - ISA 2006 Listener 1 - OWA Rule - Internal IP - Cisco FW NAT IP1 - OWA
OA:
Public IP2 - Cisco FW - External IP2 - ISA 2006 Listener 2 - OA Rule - Internal IP - Cisco FW NAT IP2 - CAS
The two published systems run on separate IP's from the internet, all the way through to the Exchange boxes, sharing only the internal IP of the ISA box.
Our OWA uses RSA SecurID, so it has a separate listener on the ISA.
The NAT on the internal Cisco Firewall runs on two sets of different IP's.
Our OWA rule works perfectly. The OA rule doesn't !
When we test OA, traffic is seen on the ISA, coming into the OA External IP, but then a new connection is initiated from the OWA External IP to the NAT IP. The connection sits there until the timeout is reached. We then see an error which correctly states the IP's in the path and says that the "connection attempt failed because the connected party did not properly respond after a period of time".
Does anyone have a step-by-step guide of how and what the ISA settings should be, as well as a method to track traffic beyond the ISA, to see if the OA request is actually reaching the CAS array ?
Thanks in advance
Having recently published our Exchange 2010 SP2 CAS Array to host Outlook Anywhere, via NTLM, I'm about to throw them across the room !
We have this scenario :-
OWA :
Public IP1 - Cisco FW - External IP1 - ISA 2006 Listener 1 - OWA Rule - Internal IP - Cisco FW NAT IP1 - OWA
OA:
Public IP2 - Cisco FW - External IP2 - ISA 2006 Listener 2 - OA Rule - Internal IP - Cisco FW NAT IP2 - CAS
The two published systems run on separate IP's from the internet, all the way through to the Exchange boxes, sharing only the internal IP of the ISA box.
Our OWA uses RSA SecurID, so it has a separate listener on the ISA.
The NAT on the internal Cisco Firewall runs on two sets of different IP's.
Our OWA rule works perfectly. The OA rule doesn't !
When we test OA, traffic is seen on the ISA, coming into the OA External IP, but then a new connection is initiated from the OWA External IP to the NAT IP. The connection sits there until the timeout is reached. We then see an error which correctly states the IP's in the path and says that the "connection attempt failed because the connected party did not properly respond after a period of time".
Does anyone have a step-by-step guide of how and what the ISA settings should be, as well as a method to track traffic beyond the ISA, to see if the OA request is actually reaching the CAS array ?
Thanks in advance
ASKER
Thanks - the upgrade to SP3 is planned.
All of the literature I have seen, tells me to point an OA publishing rule at my CAS address - the logical address of my NLB. The OWA and CAS have different host names, and the CAS array has an entry on the correct DNS servers, so it resolves correctly across the firewalls involved. We do our OWA this way and it is fine.
Are you saying that I need a new DNS entry, specifically for OA, which points to the logical IP of my NLB ?
Any more suggestions regarding the rule setup ?
All of the literature I have seen, tells me to point an OA publishing rule at my CAS address - the logical address of my NLB. The OWA and CAS have different host names, and the CAS array has an entry on the correct DNS servers, so it resolves correctly across the firewalls involved. We do our OWA this way and it is fine.
Are you saying that I need a new DNS entry, specifically for OA, which points to the logical IP of my NLB ?
Any more suggestions regarding the rule setup ?
The CAS address and the CAS array are two different things.
You can use the same IP address, but the DNS entries should be different.
Similarly the CAS Array address does not and should not be on your SSL certificate - it should be used exclusively for the CAS array functionality.
Simon.
You can use the same IP address, but the DNS entries should be different.
Similarly the CAS Array address does not and should not be on your SSL certificate - it should be used exclusively for the CAS array functionality.
Simon.
ASKER
I think you're off on a tangent here. I need to figure out what settings my ISA publishing rule should have, more than worry about my CAS array or IP.
The certificates are ok and my NAT is fine. The problem - as I see it - is the ISA rule.
Has anyone else published OA or ActiveSync under the same setup as we have ?
The certificates are ok and my NAT is fine. The problem - as I see it - is the ISA rule.
Has anyone else published OA or ActiveSync under the same setup as we have ?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
Managed to resolve the situation. Reset SSL certificates and reconfigured ISA Listeners. In the end it just started working.
1. Exchange 2010 SP2 is no longer supported. Went end of life a few months ago, so should be upgraded.
2. You have mentioned the CAS Array. The CAS array is an INTERNAL name only, it should not resolve externally. It is for INTERNAL MAPI traffic only. If you have used the same host name for anything else (OWA, Outlook Anywhere etc) then you will expect problems because the client gets confused.
Simon.