Log in



Archive for October 15th, 2009

Beware of the Snow Leopard.

October 15th, 2009 by Scott

I had a call with a client yesterday regarding their Exchange 2007 mailbox server crashing with Dr. Watson errors. Their deployment is a Single Copy Cluster and to make matters worse when the error was occurring the cluster would not fail over.

This problem started for them when Mac users upgraded to the newest version of Snow Leopard. Based on conversations I had with them I’m guessing Snow Leopard was causing a high level of traffic to the Exchange 2007 mailbox server and caused the server to fail. I have actually seen this as well when an Outlook plug-in goes rouge and causes a high level of RPC connections to the mailbox server. The mailbox server in essence becomes unresponsive and users can not connect.

Asking about their updates they had not updated the Exchange servers with any Exchange roll ups for some time. I advised them to update their servers, which they did by deploying RU 9. They also asked Mac users not to use Snow Leopard for awhile until they get this thing out of control as their mailbox server was crashing every 20 minutes give or take. 

As a note the admin there said he was also familar with a another Exchange admin having the same experience.  This may not be directly related to Snow Leopard and Exchange not being updated but there is a chance. 

I’m thinking the RU9 will help, but I’ll let you know if the problem continues!

KB974571 Breaks Office Communication Server!

October 15th, 2009 by Scott

This morning when I went to log into Microsoft Office Communicator from a client I failed to sign in.  Given that this worked yesterday without any problem I couldn’t figure out what was going on so I pinged our local admin to let him know OCS was not working.  Long story short it appears that KB974571, which our Front End and Edge Server Downloaded last night kills OCS.  Our System Admin had to remove the update and everything went back to working…..

So, be wary of the KB974571.  More info can be found here: http://support.microsoft.com/kb/974571 and it looks like MS is already aware of the situation. 

 

 

New database size limit recommendation in Exchange 2010 using Database Availability Groups

October 15th, 2009 by Scott

While reviewing some of the new changes made to Exchange 2010 I came across an interesting article regarding Database Availability Groups (DAGs) and the recommended database size limits.  Microsoft has done away with Storage Groups in Exchange 2010 and have gone directly to simply databases, which gives us the highly available architecture of DAGs.  A database availability group (DAG) is the base component of the high availability and site resilience framework that is built into Exchange 2010. A DAG is a group of up to 16 Mailbox servers that host a set of databases and provide automatic database-level recovery from failures that affect individual servers or databases.   System administrators can leverage multiple mailbox servers and replicate a mailbox database to a maximum of 16 mailbox servers, as 16 servers is the maximum number of servers in a DAG.  In the event of an issue on one of the mailbox servers an Administrator can move a user’s mailbox to another database in the DAG.  This process is seamless and requires about 30 seconds.  During the transferring of the user’s mailbox to another server in the DAG the users experience no outage! 

Now, enough with the explanation, in Exchange 2010 Microsoft has suggested a maximum database size of 2TB!  The explanation:  With the significant core improvements made to Exchange 2010, the maximum recommended mailbox database size has increased from 200 GB in Exchange 2007 to 2 TB in Exchange 2010. Supporting these much larger databases means moving away from legacy recovery mechanisms, such as backup and restore, and moving to newer, faster forms of protection, such as data replication and server redundancy.  This is where the DAGs come in play.  As I mentioned above, a database can be replicated to 15 other servers in a DAG.  Keep in mind the max servers in a DAG is 16. 

I was able to find out this information and more at: http://technet.microsoft.com/en-us/library/dd638137(EXCHG.140).aspx

Microsoft.Exchange.Data.Storage.ConnectionFailedTransientException! OWA NOT Working!

October 15th, 2009 by Scott

The Setup:

Exchange 2007, Active Directory Empty Forest root with two Children domains DomainA and DomainB, Active Directory Sites and Services had both children domains in the same Site.

The problem:

Since the two domains were in the same AD site I deployed a SCC Mailbox Cluster in DomainA with two redundant Client Access Servers.  In DomainB I deployed a Single Mailbox Server.  Keep in mind, while testing this in the Lab everything worked without a problem, Mailbox Access and OWA.  OWA was really the big thing here since users in DomainB only access their mailbox via OWA.  When deploying this solution in production I received the error “Microsoft.Exchange.Data.Storage.ConnectionFailedTransientException” in the IE Window.  When accessing the mailbox on the mailbox server in DomainB using Outlook there were no problems!  Again, accessing via OWA the error message Microsoft.Exchange.Data.Storage.ConnectionFailedTransientException kept coming up.  This error message basically meant that OWA could not establish a connection to the backend mailbox server in DomainB.  This made absolutely no sense what so ever considering this deployment worked in my lab! 

So, after running some of the tests in Exchange 2007 and not being able to resolve the issue I contacted Microsoft Support.  Long story short, it turns out that the server administrator who build the mailbox server built it in DomainA.  Once he had the server built he then joined the mailbox server to DomainB where it became the mailbox server for DomainB.  As I’m sure we all know when a server is joined to a domain there is an object created in that domain that relates to the server.  What was occurring is that when I went to OWA and tried to access a mailbox in DomainB the CAS servers in DomainA were locating the mailbox server in DomainA.  The reason?  Well, since the Server Administrator created the mailbox server for DomainB in DomainA and then joined it to DomainB there was an active directory object for the mailbox server which never got deleted.  The CASs were trying to query a machine that no longer existed in their Active Directory Domain.

 

The Fix:

After spending three days on the phone with Microsoft Support trouble shooting the issue the resolution was simple.  Yup, you guessed it, delete the old computer object in DomainA. 

Once the computer object was removed from DomainA OWA worked again for users in DomainB.  Luckily I was still in test mode at that time and no users were impacted.  The unfortunate thing is I didn’t realize the server was originally build in DomainA.  I’m wondering how many people out there have made this mistake…..

  • You are currently browsing the Scott Feltmann's Blog blog archives for the day Thursday, October 15th, 2009.