Archives of the TeradataForum
Message Posted: Wed, 13 Jul 2016 @ 19:31:22 GMT
<-- Anonymously Posted: Wednesday, July 13, 2016 14:30 -->
this is my first posting so please be gentle with me :-)
I'm trying to find a way to have Teradata utilities send a "keep alive" more frequently than every 5 mins.
Has anyone seen the same issues as detailed below?
Our Informatica load jobs (mostly Fastloads) are taking over 20 mins and are being timed out by the firewall/load balancer (F5 LTM).
Our Teradata is behind our F5 firewall (AFM) and load balancer (LTM) forwarding VIP. The LTM has a timeout of 5 mins
We have been advised that Teradata's CLIv2 has a hard coded keep alive poll of 5 mins. This results is any connections seen as idle for 5 mins (as Fastload parent sessions often are) are dropped.
Teradata doesn't know this has happened and the parent session sends a response once the data is applied, and then waits for a reply, which never arrives.
The implications of increasing the F5 timeout are huge because we would need to set it to 15 mins (practice is to allow for at least one and usually two failed tests before dropping a connection). The extra traffic on the rest of the network would cause failures on other core systems (we've already seen this).
I'd welcome any response and any questions - I'll do my best to answer (my role is Teradata DBA but I could write my experience of firewalls on the back of a postage stamp with a thick piece of chalk!)
Thanks for your patience!
|Copyright 2016 - All Rights Reserved|
|Last Modified: 24 Jul 2020|