On Feb 5,  2:11pm, Brett Walker (408)256-0265 wrote:
> Subject: Re: ANS4017E Session rejected: TCP/IP communications failure
> >> One other thing that you could try would be to reduce the size of
> >> the TCPWindowsize and TCPBuffsize parameters.  In our quest for
performance
> >> we tend to bump these values up to their maximum, but on some networks
this
> >> may be detrimental.
> >>
> >> Just a thought ;-)
> >>
> >> Brett Walker
> >> ADSM Development
> >>-- End of excerpt from Brett Walker (408)256-0265
> >
> >Brett-- if the window and buffers are too large, don't you see "out of
memory"
> >problems on the
> >client first ???  I did try massaging these numbers many times for one
> >particular client who
> >always got tcpip errors.... but it never seemed to help.
> >
> >susie McClure
> >Rice University
>
> If you try to bump them up WAY high, maybe.  But you might see problems
> if you had them only at 64K.
>
> I know that MacTCP had problems (I think it still does) where if there
> were packets unacknowledged, it would retransmit all the packets in the
> tcp window.  The bigger the window size, the more likely it would have
> to retransmit a lot of packets.  This could cause intermittent drops.
>
> Brett Walker
> ADSM Development
>-- End of excerpt from Brett Walker (408)256-0265

Brett- AAAAHHHH, the unit in question was a MAC with MACTCP. I now have the
windows & buffers set just a little over the default, have increased
commitimeout to 10 minutes and have set commandretries way up for this user....
It is a very severely overloaded system, very slow.... and now, I actually
successfully complete a scheduled incremental about 1 out of 4-5 evenings.
Most Macs I back up once a week... this one is set to back up every night in
hopes that I just might be successful once in a while.

There's got to be a better way.....

Susie McClure
Rice University