I did this today. I installed sme7, added 10 users, sent email from each of the 10 users to at least 3 other users, yum update, post-upgrade; reboot, checked server-manager, users fine. Then I logged in to webmail again as each user, no issues. Downloaded the RPMS from my contribs site, checked the md5sums and they matched. Installed the 10 RPMS you should have:
e-smith-horde-1.13.0-15.noarch.rpm
imp-h3-4.1.3-00sme01.noarch.rpm
e-smith-imp-1.13.0-07.noarch.rpm
ingo-h3-1.1.2-00sme01.noarch.rpm
smeserver-kronolith-1.1-5.noarch.rpm
e-smith-ingo-1.1-5.noarch.rpm
kronolith-h3-2.1.4-00sme01.noarch.rpm
turba-h3-2.1.3-00sme01.noarch.rpm
e-smith-turba-1.7.0-08.noarch.rpm
horde-3.1.3-00sme01.noarch.rpm
post-upgrade and reboot. Checked server-manager, no issues with users, added users free/busy information via server-manager (this must be done in order to be able to query a users calendar), logged in as users, created default identities. Opened Kronolith and added events and invited users.
You error about a horde VFS tells me that either you had an error when doing the post-upgrade, or there may be a templates-custom fragment.
Your /var/log/messages should show any of these errors. If you don't have it, do a signal-event post-upgrade again and then browse your log before you do the reboot.
John