X-no-Archive:yes
OK, so you are going from the ftp client on the 3000 to NT servers running
NT's IIS ftp server. And it's slow. Well, from the 3000, you can try some
diagnostic pings, and see what kind of response times you get. I believe you
can set a 2KB packet size. Try that, for maybe a dozen or two dozen packets.
If response time tanks, or fluctuates wildly, then it's likely the
connectivity between the two points, which is not unheard of between unlike
platforms, and especially thru firewalls.

Are you doing this in session (CS) or batch (DS)? You can set the batch jobs
to use PRI=CS. Have you increased your outbound memory buffer pool to
correspond to any increase above the default memory size on the box? I got
that tip from HPRC, and they were very helpful when we were getting SOCKERRs
(64 and 67) in our ftp jobs. Are you on the latest patches? HPRC will ask
you that, and I never regretted a network patch (although I have read some
war stories here). HPRC also knows where to look for logging information
that can indicate problems, insufficient resources, etc.

You say that you get 75 times that with the Reflections FTP client. Is this
sending to your 3000, or to those NT servers? You can use the smbclient on
the 3000 to do a put to a share on the NT boxes if their admins will let
you, or if one will at least let you try, to see if it is an ftp problem or
a network problem. Is this problem completely consistent? At all times of
day, to all NT boxes? Do you have an in-house NT ftp server that you can
test against? How busy is the 3000 when this happens?

These network problems can be like trying to nail Jell-O to a tree. You may
actually have more than one problem! You may find and solve one problem, and
only experience partial relief.

Greg Stigers
http://www.cgiusa.com