Instance TCP Port Misconfiguration On Cluster Leads To Unrecorverable Situation....

Hi all. I have ran into a annoying situatin on sql cluster. Config is:
w2k3 EE x64 r2 - two nodes
sql 2005 x64 standart v2153
2instances, active-active cluster.
While changing tcp port configuration of one instance (using configuration manager) I have acidentaly created that setting:
- All IP dynamic port and All IP port was set to same number while separate IPs have enabled dynamic ports and does not have static port set.
This situation leads to unability of starting the instance anymore. There is no need wide explanation, its quite clear - while starting instace, it was configured with static port and with dynamic port with the same number, so port conflict error takes place. Problem was how to escape from this situation. On non clustered sql is quite easy, because these settings are stored in registry and also can be changed thorough Configuration manager, easy solution. Nor on the cluster.. Modifying the registry even on both nodes was unefective, every try to start instance fails with port conflict error and bad config was saved to registry again. I wasnt able to find where this config resides and change it to corect values. I have revealed only that it is not stored in instance master DB (restoring it from backup). Looks like its held somehow by cluster service itself, but I didnt find where it should be and how to edit it. So the last possible option take place (time was against me) - reinstalation of whole clustered instance. Nice example of how one click can create you a 2 days of work... Iam still interested if there is a smooth and easy way out of this situatin:-) So if anyone knows..... If there is not, I realy warn everybody.

DAvid