Yes, it's a possibility.  Maybe even a probability.  I'll see if I can
re-create the problem and will then rerun this to make sure.

Gary

>>> Stan Sieler <[log in to unmask]> 10/18/96 09:51am >>>
Hi,

Any chance that you've restarted your machine since the last time
the problem was seen?

> Here's the results on our 968.  All of the faces are still smiling.
Except
> for ours, of course, when we get the incomplete WWW transfers :(
>
> THU, OCT 17, 1996,  9:36:09 AM
>  Item  Subsystem   Name   G/N  Description      Used   High    Max
> ____________________________________________________________________
>
>   1    NS XPORT  CP_POOL_ (G)  Control Buf Pool    0      1    300 :)
>   2    NS XPORT  1536____ (G)  Inbound Buf Pool   38     66    170 :)
>   3    NS XPORT  WHITMAN_ (N)  Outbnd  Buf Pool    0     22    128 :)
>   4    NS XPORT  WHITLOOP (N)  Outbnd  Buf Pool    0      1    128 :)
>   5    NS XPORT  WHITMAN_ (N)  StrFwd  Buf Pool    0      0     20 :)
>   6    NS XPORT  UDP      (G)  GProt  Msg  Pool   50    N/A    512 :)
>   7    NS XPORT  PXP      (G)  GProt  Msg  Pool    1    N/A    660 :)
>   8    NS XPORT  PROBE    (G)  CM Prot Msg Pool   17    N/A    678 :)
>   9    NS XPORT  IP_NI    (G)  IP-NI  Msg  Pool   50    N/A   2048 :)
>  10    NS XPORT  IP_NI    (G)  IP-NI  Msg  Pool   50    N/A   2048 :)
>  11    NS XPORT           (G)  Node  Name Cache    0      7    360 :)
>
> >>> Jeff Kell <[log in to unmask]> 10/17/96 08:19am >>>
> [snip]
> (1) Does this occur with a large HTML page, cgi output, or both?
> (2) Does the browser report an error?
> (3) Have you checked your IP buffer pools to see if they have been
> exhausted?  This is my strongest suspicion.  To check this:
>
>    :nettool.net.sys "resource;display;quit"
>
> If any of the "high" marks meet or approach their "max" this could be
> the problem.  [Trivia note:  Only last year did I discover that the
> output of this display has "smileys" [ :) ] at the end until you get
> close to the max, then they change to [ :( ] frowns!].
> (4) Can this behavior be duplicated with another BSD sockets example?
>
> Given all of the file transfer functions (FTP, WRQ, DSCOPY, etc) we
> should have hit this long ago if it were a simple buffering problem, but
> they all use NetIPC.  I can't think of an immediate example of a BSD
> example that would do such large transfers other than httpd.
>
> Jeff Kell <[log in to unmask]>
>