Upcoming version woes

I realize that there are going to be lots of complaints, and I hope 
you'll find the time to answer these questions.

All of my servers are centos 7 with epel repo.  It's time for our 
monthly maintenance window. Last updates were in April.

Clients and servers have been running:

rdiff-backup.x86_64 1.2.8-13.el7                   @epel

Two test servers were updated to 2.0.0 about 30 minutes ago, and crash 
now with errors and tracebacks:

Exception '('object.__new__(X): X is not a type object (classobj)', 
<function _reconstructor at 0x7f5d6b14fa28>, (<class 
rdiff_backup.connection.ConnectionRequest at 0x7f5d5f09aa78>, <type 
'object'>, None))' raised of class '<type 'exceptions.TypeError'>':

culminating in:

  This message may also be displayed if the remote
version of rdiff-backup is quite different from the local version (2.0.0).

which of course it is.

So do I update the backups server to 2.0.0 before updating clients? Can 
I be confident of having backups after I upgrade everything?

Should I version lock all our servers to 1.2.8?

Maintenance window is at 5 AM ET tomorrow.

Also, since the 1.2.8 packages are no longer in the repo, I can't 
downgrade the two servers that I upgraded, so I will have no backups on 
them tonight. Any suggestions?


Thanks,

Ed Greenberg