.... where to begin?

One of our slave NW6.5SP1, BM3.8SP1 servers moved to a new office at the
weekend, and got a new public IP address. I'd not yet migrated this
server to the "new" IPsec BM3.8 VPN, so it was still running the old
fashioned way.

To get things up and running as quickly as possible, I went to do the
old-style VPN configure - VPNCFG, MINFO.VPN, SINFO.VPN etc. However:

- When I edit the IP details in the slave's VPNCFG, the new address is
accepted - but if I exit VPNCFG and go back in, the old address is still
there. The SINFO.VPN that's generated would appear to have the correct,
new address (it shows correctly when I plug it into NWAdmin), so is this
just a cosmetic thing that would correct itself if/when the VPN comes
back up?

- As soon as I REINITIALIZE SYSTEM on the slave, I get this error:

" 7-07-2004 1:57:49 pm: TCPIP-6.64-43
30:1.9 Could not make WAN connection to VPTUNNEL@217-37-184-226
due to CSL failure (1).

Needless to say, the VPN never comes up.

If I use CALLMGR.NLM to force the call up, the status is "disconnected",
and F2 tells me that there may not be enough memory to connect. There's
loads of memory - server has 4Gb, 70% NSS cache and nothing apart from
BorderManager and a few other "standard" NLMs loaded.

- If I try and get the audit log for ANY of the slave servers (including
a couple that are working fine!) in the master's details in NWAdmin,
this appears:

"The vpn member was not found for the requested ip address
Exiting vpn configuration program"

I am losing the will to live with this one; I have considered biting the
bullet and moving to the IPSec VPN, but the thought of having to
configure iManager to work on another server makes me want to cry
(having set up a dozen or so NW6.5 servers, I've never had iManager
work "out-of-the-box"). Can someone save my sanity and help me?