I'm a junior sysadmin for an IT contractor that helps small, local government agencies, like little towns and the like. One of our clients, a public library with ~ 50 staff users, was recently migrated from Server 2003 Standard to Server 2008 R2 Standard in a very short timeframe; our senior employee, the only network engineer, had suddenly put in his two weeks notice, so management pushed him to do this project before quitting. A bit hasty on management's part? Perhaps. Could we do anything about that? Nope. Do I have to fix this all by myself? Pretty much.

The network is set up like this:

a) 50ish staff workstations, all running Vista Business SP2. All staff use MS Outlook, which uses RPC-over-HTTPS ("Outlook Anywhere") for cached Exchange access to an offsite location.

b) One new (virtualized) Server 2008 R2 Standard instance, running atop a Server 2008 R2 host via Hyper-V. The VM is the domain's DC, and also the site's one and only file server. Let's call that VM "NEWBOX".

c) One old physical Server 2003 Standard server, running the same roles. Let's call it "OLDBOX". It's still on the network and accessible, but it's been demoted, and its shares have been disabled. No data has been deleted.

c) Gigabit Ethernet everywhere. The organization's only has one domain, and it did not change during the migration.

d) Most users were set up for a combo of redirected folders + offline files, but some older employees who had been with the organization a long time are still on roaming profiles.

To sum up: the servers in question handle user accounts and files, nothing else (eg, no TS, no mail, no IIS, etc.)

I have two major problems I'm hoping you can help me with:

1) Even though all domain users have had their redirected folders moved to the new server, and loggin in to their workstations and testing confirms that the Documents/Music/Whatever folders point to the new paths, it appears some users (not laptops or anything either!) had been working offline from OLDBOX for a long time, and nobody realized it. Here's the ugly implication: a bunch of their data now lives only in their CSC folders, because they can't access the share on OLDBOX and sync with it finally. How do I get this data out of those CSC folders, and onto NEWBOX?

2) What's the best way to migrate roaming profile users to non-roaming ones, without losing vital data like documents, any lingering PSTs, etc?

Things I've thought about trying:

For problem 1:

a) Reenable the documents share on OLDBOX, force an Offline Files sync for ALL domain users, then copy OLDBOX's share's data to the equivalent share on NEWBOX. Reinitialize the Offline Files cache for every user.

With this:

How do I safely force a domain-wide Offline Files sync?

Could I lose data by reenabling the share on OLDBOX and forcing the sync?

Afterwards, how can I reinitialize the Offline Files cache for every user, without doing it manually, workstation by workstation?

b) Determine which users have unsynced changes to OLDBOX (again, how?), search each user's CSC folder domain-wide via workstation admin shares, and grab the unsynched data. Reinitialize the Offline Files cache for every user.

With this:

How can I detect which users have unsynched changes with a script?

How can I search each user's CSC folder, when the ownership and permissions set for CSC folders are so restrictive?

Again, afterwards, how can I reinitialize the Offline Files cache for every user, without doing it manually, workstation by workstation?

c) Manually visit each workstation, copy the contents of the CSC folder, and manually copy that data onto NEWBOX. Reinitialize the Offline Files cache for every user.

With this:

Again, how do I 'break into' the CSC folder and get to its data? As an experiment, I took one workstation's HD offsite, imaged it for safety, and then tried the following with one of our shop PCs, after attaching the drive: grant myself full control of the folder (failed), grant myself ownership of the folder (failed), run chkdsk on the whole drive to make sure nothing's messed up (all OK), try to take full control of the entire drive (failed), try to take ownership of the entire drive (failed) MS KB articles and Googling around suggests there's a utility called CSCCMD that's meant for this exact scenario...but it looks like it's available for XP, not Vista, no?

Again, afterwards, how can I reinitialize the Offline Files cache for every user, without doing it manually, workstation by workstation?

For problem 2:

a) Figure out which users are on roaming profiles, and where their profiles 'live' on the server. Create new folders for them in the redirected folders repository, migrate existing data, and disable the roaming.

With this:

Finding out who's roaming isn't hard. But what's the best way to disable the roaming itself? In AD Users and Computers, or on each user's workstation? Doing it centrally on the server seems more efficient; that said, all of the KB research I've done turns up articles on how to go from local to roaming, not the other way around, so I don't have good documentation on this.

In closing: we have good backups of NEWBOX and OLDBOX, but not of the workstations themselves, so anything drastic on the client side would need imaging and testing for safety.

Thanks for reading along this far! Hopefully you can help me dig us out of this mess.