Most of the advantages of going to TSP 1.0.0.28 and CM 3.0(7) are bug fixes, but there is also support for simple fail over scenarios (described in another post out here last week).
3.0(7h) was just the interim build the lab guys were testing the fail over stuff on, it's not an available build. The differences between 3.0(7) and 3.0(7h) should not be significant as far as the fail over testing was concerned.
Anyway, here's a short list of the bugs addressed in 28/3.0(7) from a previous post. The release notes for 3.0(7) have lots more information on the CM specific fixes/changes.
===============================================
Fixed in TSP 1.0.0.28
· A call that is blind transferred within a CallManager cluster to a subscriber's phone will not return successfully on a FWD/Busy or FWD/RNA back into Unity.
· When attempting any supervised transfer, Unity fails to dial the called phone and drops the inbound caller immediately at the moment of transfer. The inbound caller is left in limbo with the CallManager default hold.
· Each time the AV-Cisco TSP does an MWI a small amount of memory is being leaked.
· Registration Warnings and Failures in CCM Event Log, Fixed in TSP 1.0.0.28
· With Automatic Reconnection enabled, physically breaking the network connection between the AV-Cisco TSP and the primary CallManager server will cause the TSP to failover, but then continually disconnect and reconnect to the failover CallManager server.
Jeff Lindborg
Unity Product Architect/Answer Monkey
Cisco Systems
jlindborg@activevoice.com
http://www.AnswerMonkey.net (new page for Unity support tools and scripts)
