X-no-Archive:yes
Are the 9000 and 3000 local to each other? Is there a lot of network traffic
between the two systems? Or, are the files themselves large? If so, are your
inbound and outbound memory buffers scaled to correspond to the amount of
memory above the system default on the 3000? For instance, if you have twice
as much memory as the default, are the buffers also doubled?

It's been years, but when we saw problems with ftp, there was a firewall
between our system and the other system. The firewall was overloaded, and so
created problems by increasing latency. In any case, HPRC worked with us,
and had an engineer on the system after an error, who was able to diagnose
and correct our issues. I commend you to HPRC.

As for work arounds, what does the dir do for you? Can you find an
alternative? Does ls exhibit the same problem? Can you live without the dir,
or perhaps examine whatever args you are giving it? I've seen a dir on a
mainframe not respond before the inactivity timeout (the amount of time a
user will be disconnected if there is no activity). Perhaps the time the dir
takes to build its file list to ftp back to your 3000 is too long (and
that's what dir does, the results are actually a small file transferred to
the client and displayed there).

Greg Stigers
http://www.cgiusa.com

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *