Cindy,

  Our TCPBuffSize is normally @ 24 and our TCPWindowSize @ 31.
  On the remote sites we went to 12 and 16, respectively.

  I also suspect that the network in these locations is simply being
overloaded and the above change was a more economical solution than
modifying our current network.  As we are doing a network upgrade in the
near future , it did not seem reasonable to waste resources with the
current network.  If this were not the case, I would have requested that
some research be done on a better resolution.  Throttling back the backups
is fine as long as we meet our window, but restores are another matter.  If
the network cannot handle a reasonable restore, then our customers need to
be aware of this fact.

I hope I have not muddied the waters.
  Dan T.

----------
> From: Cindy Cannam <CCannam AT GENAM DOT COM>
> From: Cindy Cannam <CCannam AT GENAM DOT COM>
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: Re: ANS4017E Session rejected: TCP/IP connection failure
> Date: Friday, January 16, 1998 3:47 PM
>
> Dan:
>
> When you say IBM suggested that you halve the values you had in your
> TCPBUFFSIZE & TCPWINDOWSIZE options in our dsm.opt file, what values did
> you originally have in place? I've seen the same below-described message
> periodically, but had been attributing them to poorly defined network
> router parameters. That now may not have been the case. Once you made the
> suggested changes, did the message stop appearing?
>
> Thanks.
>
> C.L.Cannam
> Storage Management
> GENAM/St. Louis, MO
>
>
>
>
> Daniel Thompson <thompsod AT USAA DOT COM> on 01/16/98 01:48:55 PM
>
> Please respond to "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
>
> To:   ADSM-L AT VM.MARIST DOT EDU
> cc:    (bcc: Cindy Cannam)
> Subject:  Re: ANS4017E Session rejected: TCP/IP connection failure
>
>
>
>
> David,
>   This may or may not help.  We had problems with some of our NT boxes
> getting intermittent problems like you describe.  IBM suggested that we
> halve the TCPBUFFSIZE & TCPWINDOWSIZE options in our dsm.opt.  Our
problems
> seemed to be with older client code, which does not match with your
> problems.
> Good luck,
>   Dan T.
> ----------
> > From: David H. Wentworth <David.Wentworth AT UCOP DOT EDU>
> > To: ADSM-L AT VM.MARIST DOT EDU
> > Subject: ANS4017E Session rejected: TCP/IP connection failure
> > Date: Friday, January 16, 1998 11:44 AM
> >
> > We've been attempting a restore for a couple of days on a NetWare
client
> > that keeps halting with the message "ANS4017E Session rejected: TCP/IP
> > connection failure." The ethernet segment has a lot of traffic and
we've
> > seen other communications problems, so I'm guessing that's the cause.
> Does
> > anyone have suggestions on how to stop the client from halting the
> restore?
> > A setting in dsm.opt perhaps? Here are the version details:
> >
> > Server: MVS Server Version 2, Release 1, Level 0.10
> > Client: Version 2, Release 1, Level 0.7
> >
> > We're using the default settings in dsm.opt and the comm method of
> TCP/IP.
> >
> > Any help will be much appreciated.
> >
> >         Dave
> >
> > +-----------------------------------------------------+
> > | David Wentworth          | University of California |
> > | Programmer/Analyst       | Office of the President  |
> > | david.wentworth AT ucop DOT edu | Information Resources    |
> > |                          |   & Communications       |
> > +-----------------------------------------------------+