I'm suspecting you'll never get the two cameras to match.
If both cameras had WB off the same then for sure you'd get them to match. After all if it were that simple why worry about WB in the camera.
The issue is that the camera does image processing in a much larger color space before it does the sampling and compression. During sampling / compression a lot of data is lost. When you apply CC in post you can get a very good result in Vegas as it also works in a very large color space however it cannot recover that which is lost. The eye will barely notice this normally, I'd suspect only someone who is trained may pick the difference. But when you cut between stuff fom two cameras that has very different CC applied to the same objects it sticks out like dog's dangly bits.
This doesn't just apply to DV25, even with 4:2:2 sampling in 8 bit it's still not perfect. I suspect that's why film DIs are scanned at 10 to 14 bit res, it gives the guy doing color grading so much more data to work with.

Perhaps one answer to this dilema is (heaven forbid) to actualy make the good camera worse to start with. OK, maybe I'm nuts BUT given how the eye works when it comes to color, it doesn't notice absolute errors much, it's relative errors that it picks up. So try this.

CC the good footage so it's the same as the bad footage. Render that out, now CC both with the same set of values to get the WB back correct. They'll both have the same loss of detail and hopefully the eye will forgive that.

Bob

Thanks guys for your help. I spent some more time last night working with the color corrector and the secondary color corrector (to fine tune the bridesmaids dresses) and the results are great! It's difficult to tell the two camera views apart! Comparing the before and after, it's almost like magic. I'll try and post some jpgs for comparison. Whoohoo, I'll be able to use the footage from the GL2 now :)

John