After 4 hours I figured this out, already reported it to IBM. You can see this occurring in the log folder for traveler under the ibm support folder in the data folder. Never knew it was there until I had to do digging for this one.
The users which were getting this error in our organization had the replicate unread marks in the mail database set to "Never". Changed it to "All Servers" and either waited about 30-60 minutes, or ended traveler and restarted. At that time they could register their phone. Note, you may also need to clear the cache on the phone. To make this easier after discovering I was using a computer exclusively to troubleshoot, they are much more forgiving!
Not sure if this is an "enhancement" or a bug. I have requested IBM give me an answer on this, but on late Friday afternoon their Traveler developers were "on an outing". I hope to here back on Monday. This is an issue exclusive to 9.01.
Personally I hate the default value of never for this setting, but I know we have many like that. Back in the early days we did not modify the mail template and new users got the never value.