Archives of the TeradataForum

Message Posted: Tue, 27 May 2003 @ 14:37:25 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Hello-

Perhaps this helps, perhaps it doesn't, but here's what I'd look at.

I'd first create a new DNS alias for one of the cops in the system, I would then create a new DSN on the client for that cop, that would limit the client's interactions to the single node when it uses taht DSN.

From there, I would either:

a. use tcpdump on the teradata node (it's in the /etc/inet/ directory, if I remember right)

or

b. put a hardware sniffer on the line between the node and it's switch connection.

At that point, I'd also isolate the traffic with some expressions (tcpdump type), limiting my captures to only the server that is having the corruption issues. If you are corrupting out of the interface, I'd turn to NCR (GSC) at that point, there are some tools that can take a look through the stack with in the box itself, but unless you are comfortable with them, it's not really a prime opportunity to learn them on your own, the way this issue sounds.

hope that is of some use,

loadc

	<Prev	Next>		<<First	<Prev	Next>	Last>>