November 20, 2005 at 7:39 am
Thanks Jeff
I have no problem in getting the body mesage back on the CE device, its just the next bit that doesn't work. The problem is that teh customer has been using an ISDN link for the last three years without any problems, teh CE devices worked perfectly, in fact they still do, if I use the old ISDN.
I have installed teh BT NET 2mbps ADSL link and installed a firewall and router, with teh radius server, but through the new connection I get this problem. When i go back to the old ISDN It works perfectly and I can't see why?
What is baffling is that the SQL side give a succeeded message and then I assume it is starting to send the database back but it just seems to go nowhere.
I will have a bash at the network monitoring next week, but I am getting pretty desperate. Do you know of anybody or company that I could hire for a day or half a day to resolve this once and for all?
Regards
Matt.
November 21, 2005 at 7:31 am
Hello,
I can't think of any particular company that you could hire. It probably depends on where you're located. If you are close enough to Milwaukee, Wisconsin for me to know anybody, I'd just offer to look at it myself.
I'm thinking that you are a very long way away, since "BT" is called "Qwest" in these parts...
Hang in there!
jg
November 21, 2005 at 12:56 pm
Sorry, I am in the UK, didn't realise where you were!!!
Anyway teh plot thickens, I have had teh telephone guys monitoring the TCP packets and anything over 1420bytes gets a time out on an extended ping.
Is there any way of adjusting the packet size that SQL sends back the database in?
Thanks
November 21, 2005 at 1:11 pm
You can try to lower the MTU on the RDA server's network interface. I'm not sure how to convince RDA to use smaller packets. If I uncover anything, I will let you know.
This is for Windows 2000...
These parameters for TCP/IP are specific to individual network adapter cards. These appear under this Registry path, where "adapter ID" refers to the Services subkey for the specific adapter card:
HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\Tcpip\Parameters\ Interfaces\[Adapter ID] MTU: Set it to equal the required MTU size in decimal (default 1500)
November 21, 2005 at 1:25 pm
Thanks I will give that a try in the morning.
November 21, 2005 at 9:17 pm
I have had the same thing happen (packet size + fragmentation) on some VPNs before...
Lowering the MTU should help fix that. You could also try to fiddle with some of the VPN options, even try a "lower" form of VPN such as PPTP to see if that works to at least give you a base to prove that SQL replication + VPNs can mix.
November 21, 2005 at 11:56 pm
Thanks Ian
Will give it a go, but can't change the security as we are using an l2tp tunnel through BT to get teh connection, I am beginning to wonder if it is ever going to work!!!!
Matt
November 22, 2005 at 7:15 am
There is an interesting and fact-filled discussion of what might be going on from our friends at Cisco. This has to do with PMTUD. (Path MTU Discovery) The MTU adjustments should be automatic, but there are things you (and others like BT) can do to make it not work.
http://www.cisco.com/en/US/tech/tk827/tk369/technologies_white_paper09186a00800d6979.shtml
Let us know how things turn out!
jg
November 22, 2005 at 11:03 am
Thanks Jeff
That should keep me going for a little while.
Matt
Viewing 9 posts - 16 through 23 (of 23 total)
You must be logged in to reply to this topic. Login to reply