exch 5.5 behind proxy 2.0 email attachment hangs persistant demand dial link
To work through this problem it may help if you use a pen a paper to draw a
picture of the domains (Site 1 and Site 2) and the link between them.
Site 1has exch 5.5 sp0, on nt server 4.0 sp3 has fake IP Address 192.168.1.2
is sitting behind a nt server 4.0 sp3 with RRAS hotfix 3 and pptp hotfix and
Proxy 2.0 and a Real IP Address which has a persistant dial on demand link
nt server 4.0 sp3 at Site 2 with RRAS hotfix 3 and pptp hotfix and real IP
that is acting as the router onto the internet and primary DNS for both
domains (Site 1 and Site 2)
Also have at seperate Exch Server 5.5 sp0 on nt server 4.0 sp3 and real IP
Address that is running proxy 2.0 and DNS for (Site 2)
Exch Server (with fake IP) at Site 1 is set to forward all email to Exch
Server (with real IP) at Site 2.
everything is perfect email goes in and out, proxy's do proxy, can ping to
anywhere and resolve DNS names.
so all the bindings between the proxy and exch server ports are correct and
pass all the tests in the MS Knowledge base that relates to this setup.
BUT when someone at Site 1 (the server behind the Proxy 2.0) sends an email
with an attachment (over about 15k in size) the Modems on the demand dial
link between Site 1 Proxy server and Site 2 NT server/router "hang" if you
try to ping across the link you get "request timed out" the only way to get
the link going again is to manually dropout one of the modems and let the
link re-establish itself.
The link will then stay up until the Site 1 Exch Server Tries to resend the
email with the attachment.
This fault only applies to email with attachments that attempt to leave Site
I have successfully sent email's with attachemts up to 14Mb in size from
external email address into Site 1
This fault is not directly related to the size of the file going down the
modem, FTP and Web transfers of large files can also get into site 1 without
This fault was originally occuring when all servers only had standard RAS
and sp3 installed, I upgraded to the most recent RRAS & hotfixes in an
attempt to get away from the fault, but the symptoms are Identical.
I have changed the registry setting at the NT Server Router to prevent the
RemoteAccess AutoDisconnect from dropping the line after 20 minutes.
Both servers that are routing/proxying have static routes set, they can ping
off-site. Site 1 can ping through Site 2 to the Internet.
The attachment settings in the Site 1 & Site 2 Exch Servers are Default i.e.
I have not loaded Exch Server 5.5 SP1 at either site, after reading the
known bug list for that service pack, it appeared to have nothing to do with
WHAT DOES WORK
If I have the Site 1 Proxy connect to the internet by dialing a UNIX based
ISP, then the fault goes away, the emails route out of site 1, into the
external UNIX ISP, from there back down the net to the Site 2 NT Router,
through that to the Site 2 Exch Server, to then be forwarded to the final
destination, back out of the Site 2 NT Router onto the Net.
The problem only seems to exist when the routing of the email occurs
directly between the NT servers running RRAS/RAS adding the UNIX ISP between
them makes the fault go away, unfortunatly this can not be a permenant
I have logged a call for this with the Microsoft Support People, they have
never heard of such a thing before and after two days on the phone can not
come up with a solution.
If anyone knows this fault or has a suggestion for it's resolution, please