At 11:06 AM 8/9/00 -0600, Steve Rasmussen wrote:
>Pramod Sokke wrote:
>>
>> I've been using LWP with the timeout feature turned on and everything works
>> fine. There are about 5 different components in our entire application that
>> I'm using LWP for and in each instance, the timeout is set to a different
>> value based on the usage(for eg, a db-related call will need more time than
>> a simple GET or POST). And in each of the components, the caller of the API
>> sets his own timeout period too. And everywhere it works perfectly fine.
>>
>> -Pramod
>
>I'd guess that you are not talking to a large number (50+) of high
profile/high traffic
>web sites that you both don't own, nor you are close to.
>My bet is that you're LWP'ing to local servers that you either own or are
physically
>close to.
>
>In my experience, that is the *only* situation where LWP timeouts aren't a
big problem
>that takes many hours/weeks to gradually fix. At least, that is my
experience - maybe your
>mileage varies?
>
>Steve
Well, that's not entirely true. We run a very high traffic website here and
while there are instances where we LWP to local servers too, there are more
instances in a single process where we LWP to remote servers which are not
in our domain. We provide certain services to our customers and each time
such a thing happens, we send back data to their servers through simple
GETs or POSTs. Some of these customers make some db operations on their
ends and hence take longer to respond back, so we set our timeouts based on
that. Works fine in every instance. And we talk to high profile sites like
CNBC.com, Quicken.com, etc.
-Pramod
