Friday, February 29, 2008

Exchange 2007 leap year bug

February 29th; what a day. I found that I could not create new users and isolated the problem to the date (it worked if you changed the system time to March 1st and stopped the windows time service). This is from MS PSS...all i can say is wow...its like y2k but stuff actually broke :-)

John,

You are absolutely right.

Symptoms of this include not being able to create mailboxes or create public stores or update email policy and results in an error “The Exchange server address list service failed to respond. This could be because of an address list or email address policy configuration error”. we found this to be a Leap Year bug and if we reset the date to either 28th Feb or 1st March it works.

We strongly recommend not to change the system time to work around the issue as this may cause undesirable results

I shall update as soon as we have some kind of solution to this.

Thursday, February 28, 2008

Custom DSNs on Exchange 2007

Tired of end users calling asking what an NDR means when its relatively clearly stated in the message body? The cmdlets set-systemmessage and new-systemmessage will allow you to customize them so they are even more clear, although somehow I think users will still call :-)

This example sets DSN text for an external 'unkown mailbox" 5.0.0 fail.

[PS] C:\>New-SystemMessage -DsnCode 5.0.0 -internal $false -Text "The intended recipient was not found. Please check the spelling and e-mail address of the recipient." -Language en

All you have to do is change the DSN code and the boolean for -internal to manipulate all the DSN's you want.

Tuesday, February 26, 2008

Generic Exchange 2007 Deployment Thoughts

I was recently asked to review a 2007 deployment plan, and the high-level feedback I had was so darn standard, I figured I'd post it. The deployment was 2 CCR clusters, 4hub transports, 4 CAS, 1 Edge and 1 SCR target. The CAS and HUB servers were to run on VMWare. There was also a 3rd party SMTP gateway.

I would definitely recommend SCR, but write and test your failover powershell scripts/commands before you need to use them. As you also probably already know, virtualized exchange servers are not supported by Microsoft, but that does not mean it can’t work. Remember with CCR and SCR DB paths have to be IDENTICAL, so remember to use mount points, not drive letters. Also, don’t use the same DB/SG names on your separate CCR clusters if they are going to target back to a single SCR box (the file paths then become identical). Have a look at pfmigrate.wsf to help with public folders if you need it.

Another question to ask is, do you really need edge if you have a 3rd party SMTP gateway. If you have something like postini or frontbridge, you might be fine with routing inbound mail right to a hub transport server from the SMTP gateway.

For your cas/hub, you can run them both on the same machine as long as you don’t plan on using NLB, so you may not need vmware if that works for you.

My favorite site is msexchangeteam.com, but again, I’m sure you’ve seen it. If not, read up on their info on CAS certificates, very different requirements on 2007 than 2003 FE servers.