Forum Discussion
Patrick_Brown_7
Nimbostratus
NimbostratusIs anyone running Elite 3E
Elite 3E is a legal sector web based financial application.
We have this solution deployed. There are five IIS web servers and a pair of F5 HA LTM boxes up front. There is a clustered SQL server at the back end.
Users of this app have been complaining that performance is bad and that running long reports sometimes results in what they call "the circle of death" (IE9 waiting). Things work fine if a user connects directly to one of the web servers. I'm wondering if this might be a persistence issue. It is setup for cookie persistence.
I'm going to deploy HttpWatch on some of the user machines. Any other ideas would be great.
--Patrick
hooleylistCirrostratus
Hi Patrick,
If you have a single pool member that can handle all of the traffic, you could disable all but that member to check for persistence issues.
As a test, you could also try switching to a FastL4 profile if you don't need any L7 options.
Else, as you say, a client trace with HTTPwatch and a tcpdump of the client and serverside legs of the traffic in/out of BIG-IP would definitely help.
Aaron
Patrick_Brown_7For anyone interested, the final solution was to create a dedicated client TCP profile with a much higher timeout value. The default is 300 seconds. I set it to 1800 and the problems went away.
--Patrick