For some years now, I've been using a TCPCLIENT stage to
move data from an HTTP server to CMS.  A couple months ago
our network was reconfigured; an iPlanet server (IIRC) was
replaced by Apache.  HTTP transfers began to stall after
a couple megabytes.

I reported this to our sysprog, who readily reproduced the
behavior, _except_ that when he enabled tracing the problem
vanished.  Timing dependent?  I got the idea that if I used
a proxy it might change the timing enough to circumvent the
problem.  That worked for the couple months, until another
reconfiguration removed the proxy I was using.  I can use
YA proxy, but it does not appear to circumvent the problem.

Has anyone else seen anything like this?  Known problem?
How can I report it if we can't trace it?  How do I find
pipelines version to report it?  Is there a library
distribution I might try instead?  Where do I get it?

Many years ago, I was using RXSOCKET before I switched to
pipelines.  I might need to find that code and dust it of.

Etc.

Thanks,
gil

Reply via email to