I'm cross-posting this from my personal blog. It is quite relevant to some of you, so I thought I would share it over here as well.
We loaded Leopard on a couple of machines yesterday. It went very well, except for two significant problems.
First, Active Directory integration is broken. It centers mostly around authentication issues. If the Mac was joined to the domain before it was upgraded, it can't log on after. Directory Utility returns a "Server can not be contacted" error. If the machine was not joined to the domain, it fails while trying to with an "unknown error" in step 3 of the bind process.
The other issue is in Safari through a Microsoft ISA 2006 proxy server. When going to an SSL website, Safari crashes after it tries to authenticate. Firefox still works.
I'm guessing that both issues are related to the re-written Kerberos engine. Our call with Apple support has been escalated to engineering, so I'll post when we get a solution.
Update: As of 4:00PM EDT on Monday Oct 29, we haven't heard anything from Apple support.
Update 2: One of my coworkers didn't have a problem joining his domain at home. The difference is that he is in AD Native Mode and has DNS/DHCP hosted on Linux at home. We're in Mixed Mode and Windows-based here at work. Still no update from Apple.
Update 3: Matt (below) had success by adding his root-level domain as a search domain in the network preference pane. It didn't help me (I had already set it), but give it a try if you're failing in step 4 of the bind. I'm still failing in step 3.
Update 4: 10.5.1 didn't seem to fix my problems, but it fixed some that others were having. Apply it if you haven't already.
Update 5: We finished out migration to Active Directory native mode on our domain today. Both of my Leopard machines will now join the domain and the Safari crashes are not happening anymore.
Update 6: 10.5.2 took all of the fidgeting out of the bind process for me. It works like a champ on all of the machines we’ve tried it on. Safari still crashes when authenticating to our ISA server in production, but not in the lab. We’re still trying to determine what could possibly be causing that issue.