Cas Array HA 2 sites

Status
Not open for further replies.
D

dragan_penkov



Hi

My current design is :

3 Dag members in primary site ( CAS,HT MBX ) using HW load balancer and casarray1 pointing to balancer.FSW different server

2 DAG members in secondary site ( CAS,HT,MBX) with casarray2 also pointing to balancer with the same IP addres as casarray1. Alternate FSW in place.

HW load balancer is configured with 3 IP addresses from primary site as priority 1 and 2 IP addreses from secondary site as priority2.

Will this be OK in case of failure of primary site as per CAS server role?

Kind Regards,
 
B

Brian Day MCITP [MVP]

You'll want to do some testing, but if the RpcClientAccessArray attribute on the DBs is casarray1, then your clients will hit casarry2 and immediately want to redirect and connect to casarray1 even if it is down.Brian Day, Overall Exchange & AD Geek
MCSA 2000/2003, CCNA
MCITP: Enterprise Messaging Administrator 2010
Microsoft MVP, Exchange Server
 
D

dragan_penkov



Hi

We have configured the 3 IP address from primery site with priority 1 so it will never go to other 2 ip adresses as long as 1 of primery site IP is up.

But when all 3 Ip addreses from primery Site are down ( site failure) then it will go to 2 ip addreses from my secondary site and that is what i am trying to accomplish. So if they hit cassarray2 and connect to casarray1 then again they will be redirected to my secondary site. Is this Ok ? I will test this for sure but i was trying to get some opinion or expiriance about this kind of settings.

Dragan Penkov MCSA,VCP
 
B

Brian Day MCITP [MVP]



You may end up in some kind of redirect loop if you use priorities on the load balancer like that.

On your mailbox databases do a powershell command like..

Get-MailboxDatabase <dbname> | fl identity,RpcClientAccessServer

What is RpcClientAccessServer set to? That is the DNS name the clients will want to connect to for MAPI if their mailbox is on that DB.

Brian Day, Overall Exchange & AD Geek
MCSA 2000/2003, CCNA
MCITP: Enterprise Messaging Administrator 2010
Microsoft MVP, Exchange Server
 
D

dragan_penkov



Hi

casarray1 10.10.10.100 - Ip adress of HW balancer

casarray2 10.10.10.100 - ip address of HW balancer

HW balancer is balancing to:

10.10.10.101 primary site priority 1

10.10.10.102 primary site priority 1

10.10.10.103 primary site priority 1

10.10.20.101 secondary site priority 2

10.10.20.102 secondary site priority 2

RPCClientAccessServer is set to casarray1 on all databases.DNS name that clients will use is casarray1.All database are ponting to casarray1 so user request will go to HW load balancer . IF any of primary site IP is up then CAS server in primary site will respond and connect the user to correct database. If all 3 IP adress in primary site are down then it will go to one of CAS servers in secondary site and make the connection.I know that i can do this with DNS change but i am trying to avoid this.Also i am trying to avoid using different load balancer for secondary site.In case of full DR in primary site ( including HW load balancer that is in primary site but different location) then my plan is to make DNS change and point casarray1 to one of secondary site CAS server.Casarray2 is not used at this moment but my plan was if I have database in future that has this value set for RPCClientAccessServer then to use the same logic as above.Majority of users are in primary site.

Dragan Penkov MCSA,VCP
 
A

Anbu Selvan



Hi,

There is an another option which you have to look at as well.

The datacenter Switch Over should give you more information on the failover and fallback in a Disaster Recovery Scenario.

DACP also provide more value add in the Exchange Infra.

Kindly browse through and it may be giving you more info.

http://technet.microsoft.com/en-us/library/dd351049.aspx
With Best Regards Anbu
 
D

dragan_penkov



Hi

I am aware of this but my question was about CAS HA design using HW balancer in case of site failure.

BR

Dragan Penkov MCSA,VCP
 
X

Xiu Zhang



Hi,

I think we have to manaully modify the RPCclientaccess value on the mailbox database when primay site is down.

If we do not have value set on RPCclientaccess, then it will always try to connect the first CAS server.If first CAS server fail, it will not try to connect to the other CAS servers.
Although we have set priority on HLB, it will not try to connect to those CAS server, if the CAS1(CAS array created in primary datacenter) is down, it will not failover to the CAS server in secondary datacenter.

Understanding RPC Client Access

http://technet.microsoft.com/en-us/library/ee332317.aspx

Regards,

Xiu
 
D

dragan_penkov



Hi

Just to clarify.

CAS1 is set as RPCclientaccess on all mailboxes.

CAS1 is DNS record pointing to HLB. All clients connecting to CAS1. In behind i am balancing CAS1 to 3 CAS servers in primary site ( priority 1) and 2 CAS servers in secondary site ( priority 2).

Clients are going to CAS1 and if all 3 CAS servers in primary site are down it will be redirected to CAS server in secondary site.

So CAS1 as per client perspective is not down even if all 3 CAS servers in primary site are down ( it will be redirected to secondary site CAS's).I am just trying to avoid changing CAS1 dns record to point to secondary site CAS server in case of CAS failure in primary site.

Are you trying to say that clients will somehow know that CAS1( all members in primary site) are down? My understading was that they will go to HLB and if they get response from some CAS server in behind everything is OK.

Regards

Dragan Penkov MCSA,VCP
 
X

Xiu Zhang



Hi,

Please understand that we use RPCclientaccess parameter to let it know that which CAS it need to use. If you do not have value on RPCclientaccess on the database, then user in this database will try to use the first installed CAS to try to find the mailbox.

For your scenario, when 3 CAS servers in primary datacenter come down, it will not try to connect the other 2 CAS in seconday datacenter. Becasue, we set RPCclientaccess value as CAS1, it will always try to find CAS in CAS1 to connect to the mailbox.

Note: When you try to connect to the mailbox, it will try to connect to GC and find the user properties and find the value on homeMDB. After that, it will try to query the legacyExchangeDN value on database. value for legacyExchangeDN should contain the CAS server/CAS array there.

Regards,

Xiu
 
D

dragan_penkov



Hi

Well i have finally tested this scenario in production and it is working.

We have marked 3 primary site CAS servers as down in our HLB and Outlook connection was restored throw CAS server in secondary site.

So even if all 3 CAS servers are down it will go to CAS server in secondary site because CAS1 is a stupid IP address of VIP in HLB and the HLB is configured when 3 ip address in primary site are down to go to CAS servers in secondary site.

Regards.

Dragan Penkov MCSA,VCP
 
J

JRVA_Q



Hi Dragan,

Thanks for posting this scenario here. We are just about to configure something similar. Just to be clear:

You have two cas arrays, with one in each site. Each cas array points to the same ip address of your HLB virtual server.

If you take down every CAS in the primary site cas array, all your clients are pointed to the secondary site via the HLB.

This scenario makes more sense than manually changing the RPCClientAccessServer parameter every time you need to failover. I suppose the only time you'd need to manually change something is if your loadbalancers were down for any reason. In that case, I'd probably change the DNS before I would rehome everyone's client access server.

Thanks for testing this.
 
D

dragan_penkov



HI

No problem. That's exactly what i was trying to explain and acomplish.

Regards.

Dragan Penkov MCSA,VCP
 
X

Xiu Zhang



Theoretically speaking, it could hard to achieve. But I do not have local lab to test the environment, since we do not have HLB. Thanks for yours sharing here. :)

Regards,

Xiu
 
T

Travellling Scott



Dragan,

I like your design to help minimize any manual steps for a site failover. What happens if the HLB dies? What kind of redundancy does that have?

Scott
 
D

dragan_penkov



Hi Scott

Sorry for late answer. In case both of HLB dies ( we have HLB in cluster) then you have to make changes manually. For example in my DR procedures i am changing the DNS Host record that is pointing to HLB VIP address to point to one of CAS servers in DR site.If you have some HLB or WNLB there then you should point to that VIP address.

BR

Dragan Penkov MCSA,VCP
 
Status
Not open for further replies.
Similar threads
Thread starter Title Forum Replies Date
L Outlook clients did not reconnect to Exchange when one CAS server in CAS array became unresponsive Exchange Server Administration 1
B CAS Array and NLB Exchange Server Administration 3
D CAS Array Question Exchange Server Administration 11
M Proper way to install Exchange 2010 SP1 on a CAS Array Exchange Server Administration 3
J Segmenting IMAP traffic from CAS Array Exchange Server Administration 8
B CAS Array and Outlook 2003 clients Exchange Server Administration 5
S CAS array in Exchange 2010 Exchange Server Administration 1
G Exchange 2010 CAS Array Exchange Server Administration 4
J Geographically redundant cas array Exchange Server Administration 15
D CAS Array Questions Exchange Server Administration 3
T Exchange 2010 CAS Array setup and lab Exchange Server Administration 18
S NLB exchange 2010 CAS Array Exchange Server Administration 4
S Access CAS Array Behind ISA 2006 Exchange Server Administration 6
C Introduce another CAS/HT server on the cas array on NLB Exchange Server Administration 7
P DNS Round Robin on CAS Array Exchange Server Administration 5
B After rebooting one of the cas array servers blackbarry stops working Exchange Server Administration 1
R cas array owa is slow Exchange Server Administration 5
T CAS Array/RPCClientAccessServer and Outlook profile experiences Exchange Server Administration 8
C Reboot a CAS array member causes Outlook 2010 clients to prompt for credentials. Using Outlook 5
K Outlook client on Exchange 2010 changes cas array name to instance-<guid> Using Outlook 21
A Why not using DAG virtual IP/fqdn for CAS array in two nodes setup? Exchange Server Administration 2
A Exchange 2010 CAS Failover from Internet Facing site to Non-Internet Facing Site - Certificate Issue Exchange Server Administration 3
C Watson Error on CAS Exchange Server Administration 1
S Exchange 2010 CAS/HT/Mailbox moved - best practice/steps for decommissioning 2007 Exchange Server Administration 3
S Load balance Autodiscover with multiple CAS servers. Exchange Server Administration 2
S Load balance Autodiscover with multiple CAS servers. Exchange Server Administration 3
M Fundamental CAS question for Exchange 2010 and 2007 Co-existence... Exchange Server Administration 7
M Exchange ActiveSync HTTP 500 Exchange CAS/HUB 2007 and Exchange 2003 BE & MBX CCR 2007 Using Outlook 1
S Exchange CAS server OAB directory not showing web.config file Exchange Server Administration 3
T RBAC Error on UM/CAS Servers Exchange Server Administration 2
R FQDN of Exchange CAS servers not in SSL cert Exchange Server Administration 2
D CAS Autodiscover using -rpcclientaccessserver Exchange Server Administration 5
D CAS Design Question Exchange Server Administration 4
D Exchange 2010 CAS at 2 different Sites Exchange Server Administration 2
T Cas server uninstall failed... Now how to reinstall? Exchange Server Administration 4
S Do you need a CAS Server in Order to Use OWA in Exchange 2010 (E14)? Exchange Server Administration 9
S Any system impact if change exchange 2010 CAS/HUB and Mailbox server IP address Exchange Server Administration 1
B Exchange 2003 OWA/OA/AS over NAT and Exchange 2010 CAS Exchange Server Administration 4
B Re: Exchange 2007 and CAS from Exchange 2010 problem Exchange Server Administration 15
M Re: Exchange 2007 and CAS from Exchange 2010 problem Exchange Server Administration 2
I Exchange 2007 and CAS from Exchange 2010 problem Exchange Server Administration 4
T ActiveSync proxy problem from Exchange 2010 CAS to Exchange 2007 CAS Exchange Server Administration 5
M CAS Issue: OWA/Outlook remote users cannot login Exchange Server Administration 1
C CAS Proxy problem Exchange Server Administration 5
K Exchange 2010 OWA redirection between 2 CAS Servers Exchange Server Administration 5
S exchange 2010 cas memory usage Exchange Server Administration 2
P Exchange 2010 SP1 Cross Site CAS connection disable Exchange Server Administration 2
R Multiple CAS - Best practice with certificate request Exchange Server Administration 4
H Outlook 2007 periodically disconnects from Exchange 2010 CAS - OWA works Using Outlook 4
C CAS Connections overview Exchange Server Administration 9

Similar threads

Top