Actually, we did get to the bottom of it. It was the upgrade as such. One of our colleagues had noticed that some of the domain names on our switches were different, so he set about standardisation.
However, he forgot that the domain names are set differently for a reason. Anyway...long story short. When we rebooted the core switch after the upgrade, the multicast routing table was cleared and because the domain name had changed, the multicast could no longer do a successful lookup.
Domain names were changed back and this resolved our problem.
Ok great. I found a bug (CSCtx78044) in SXJ2 where IPmc is delayed 6-8 seconds on SXJ1 so hopefully after upgrading to SXJ3 this will resolve the issue for us. Downtime tonight and I will post the results.
We have two 6509s (sup720-3B) with SXF6 and planning to upgrade them to SXJ2 this satuday? did you notice any issues other than multicast in SXJ2