COXMAR
asked on
Simulate a Failover in Exchange 2010
How do I test my Failover Cluster in Exchange 2010. I have 2 servers with the following roles on both servers. CAS/HUB/MB. I have a CAS Array set up as well as a DAG. I have successfully made copies of the DB. Today I turned off SVR 1 in the hopes that failover would work. It did not. The Database copy took a bit, but it finally mounted on SVR 2. I can open up Outlook 2007 for my test user account, it is kind of slow, but it opens and that is all I can do. I cannot send or receive email to that test account. My Exchange is still in its testing stage and I have successfully created a couple of test accounts and mail flow does work for inbound and outbound traffic, but did not work during my hard failure test. What am I missing? I have been unable to find steps to simulate a hard (not planned) failure and a soft (planned) failure.
You have only 2 servers with all roles inside both. How are you balancing CAS conections? Did you configure correctly CAS server on DAG databases?
It is working as per design. As you mentioned, db failed over to server 2 and you were able to connect mailbox. So DAG is working.
What error are you getting, when you are trying to send an email. Remember, DAG is only for MBX. For CAS you need HLB, for HUB you need multiple hub servers.
I assume your FSW is on separate box.
What error are you getting, when you are trying to send an email. Remember, DAG is only for MBX. For CAS you need HLB, for HUB you need multiple hub servers.
I assume your FSW is on separate box.
The problem here is the fact that you have all roles on each server.
Your CAS Array and DAG cannot run on the same server.
So the best option here would be to either move your mailbox role to a different server and have the others reside on the current ones or vise versa
Your CAS Array and DAG cannot run on the same server.
So the best option here would be to either move your mailbox role to a different server and have the others reside on the current ones or vise versa
@Hendrik
CAS Array is just a object in AD, read this.
http://blogs.technet.com/b/exchange/archive/2012/03/23/demystifying-the-cas-array-object-part-1.aspx
CAS Array is just a object in AD, read this.
http://blogs.technet.com/b/exchange/archive/2012/03/23/demystifying-the-cas-array-object-part-1.aspx
Ok let me rephrase. You would normally use Windows NLB to load balance the 2 and you cannot use Windows NLB and DAG on the same server.
When I said CAS Array I was referring to the NLB sertup. :)
When I said CAS Array I was referring to the NLB sertup. :)
ASKER
Thanks for all the Comments, we are not using a NLB you cannot use that in my configuration it will not work with a DAG. The FSW is on another Member Server as it has to be in my configuration, yes proper permission (ETS) was added to the local Admin account. For AMIT - No error, email just stays in the outbox and any email that is sent to that test account does not get delivered. I turned on SVR 1 before leaving work and this morning all messages Sent and Received were delivered within 15 minutes after turning SVR 1 back on. There has to be a way to successfully test a hard and soft failure, I rather know that works now while still in the testing phase before it becomes a live Server.
Both server are in same site?
ASKER
AMIT - Yes, both servers are in the Same Site. I am using DNS as my sort of Load Balancer. I have my A record for the CAS Array pointing to both servers. I do not have DNS round robin enable for the two servers, I am thinking that I need to do that, but I am not sure. I will be reading up on that today. The Servers are Windows 2008 R2 Virtual. Small User base - which is why I opted for 2 multi-role Servers and the only reason I have 2 servers is more for redundancy in case I experience issues with one of the servers.
For CAS you need a load balancer. Then only you can achieve High Availability. As far as, your issue, it is not related to CAS it is related to HUB role. Can you send an email to test account itself and tell me the result.
We have had two servers for Exchange 2010 for years in two different sites. All roles on h82bu other servers. We would manually modify DNS for the CAS array to point to the right server, but that was the only manual step required for failover.
The best thing to do it try it, and troubleshoot all the issues you find. Make sure both servers have a send connector.
The best thing to do it try it, and troubleshoot all the issues you find. Make sure both servers have a send connector.
ASKER
Good News and another question. So the failover actually worked because the DB from SVR1 that I shutdown is still on SVR 2 and mail is flowing. The mailbox that was original on SVR1 and that moved to SRV 2 as it should after I shutdown SVR 1, but could not receive or send emails was because when I first set up the Test MBX'S I had not created the CAS Array as of yet. According to documentation it is best to create your CAS Array before you create or move any mailboxes, I knew this after the fact. Anyway I needed and just did a repair on that mailbox in outlook and mail flows incoming and outgoing with no problem on the SVR 2. My question is how to I get the MB back on it's original SVR 1? It should not be a move command, is there an activate command or something that I need to run on the EMS? I plan on testing all again, but after I figure out how to get the DB back to it's original Server.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
Thanks - Activated the original and it is back where it needs to be. Thanks for the article.