RE: "Cannot Connect" FTP -- MIDRANGE-L

It's spooky.  We started having the same problem the same day you reported
this one.  We have an RS/6000 acting as an integration hub running Mercator
to perform FTP to multiple hosts, including two AS/400's.  Thousands of
files flow through Mercator daily to our AS/400's via FTP.  The RS/6000 was
receiving intermittent FTP errors when sending to the AS/400.  The files
would arrive normally when resent.  The frequency of errors built to an
intolerable level last week.

Our Systems Administrator worked with IBM support and eventually decided to
increase the AS/400's TCP send and receive buffer sizes.  Ours had been
defaulted to 8192 (8K), and IBM recommended 65536 (64K).  Resources are
allocated in 8K blocks.

GO CFGTCP and 3. Change TCP/IP attributes 

or

CHGTCPA

The change will require that you end and restart TCP/IP.

Give it a look...

-Jim

James Damato
Manager - Technical Administration
Dollar General Corporation
jdamato@dollargeneral.com


-----Original Message-----
From: D.BALE@handleman.com [mailto:D.BALE@handleman.com]
Sent: Wednesday, May 09, 2001 10:50 AM
To: MIDRANGE-L@midrange.com
Subject: "Cannot Connect" FTP


In our environment, we have a mainframe that FTPs hundreds of files daily to
our branch AS/400s.

Invariably and unpredictably, a mainframe FTP job will abend with "Cannot
Connect".  A few minutes later, another mainframe FTP job will run
successfully.

I've been asked to look at the AS/400 side to see if there's anything we can
do to keep this from happening.  This is NOT my area of expertise.  What
I've
noticed is that our branch boxes have numerous QTFTPnnnnn jobs, for example,
our Indianapolis box has 13 of these jobs "active".  When I prompt CHGNETA
on
this box, the "Number of initial servers" is set to 5.  I presume (awaiting
correction from the list) that somewhere along the line since the last IPL
in
October, we had 13 FTP server jobs actively transferring files from the
mainframe to the AS/400.

So, I have been asked to consider changing the "Number of initial servers"
to
the maximum of 20, with the goal of minimizing the "Cannot Connect"
mainframe
abends.  This doesn't make complete sense to me, but I'd be willing to try
it
if there's no downside to doing so.  I presume that if a server job is not
in
use, eventually it gets paged out of memory with no hit on performance.

Or are we slaying the wrong dragon?

- Dan
Dan Bale says "BAN DALE!"
IT - AS/400
Handleman Company
248-362-4400  Ext. 4952
D.Bale@Handleman.com
  Quiquid latine dictum sit altum viditur.
  (Whatever is said in Latin seems profound.)
+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator:
david@midrange.com
+---
+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---