Zmodem upload problems

[Harvey R Moran]

In article <1989Dec21.200309.17636 at ux1.cso.uiuc.edu> garyf at mehlville.ncsa.uiuc.edu (Gary Faulkner) writes:
>In article <80 at ucunix.SAN.UC.EDU>, rainwatr at ucunix.san.uc.edu (Don
>Rainwater) writes:
>> 	I'm able to download (sometimes very large) files from Ultrix
>> without any trouble, but uploads don't even get past the first byte.
>> The error I get is "ZRPOS=0".  I have the same problem trying to upload
>> from an IBM/PC-AT compatible running DSZ, so I think we can eliminate
>> the Mac from the list of suspects.
>> 
>> 	Has anyone been able to successfully upload files to an Ultrix
>> (or other Unix) host?  Any help would be greatly appreciated.  Kermit is
>> just too slow.
>
>Yep, we have the same problem with our ultrix and our sun hosts using
>rzsz, again for uploads only.  I am taking the liberty of cross posting

    sz/rz works fine here for both upload and download.  Points of
comparison:  zmdm1.43, VAX-750, Ultrix 3.1, DECSERVER, IDX line mux,
Vadic modem, 1200 (yeah I know, but that all I've got to work with)
<--> Telebit T1000 modem, Microport System V/AT 2.3, 8 Mhz '286
AT-clone, xcomm PD communications package, zmdm1.43 transport.

   Though I've used it only twice, the same AT up/downloads with a '386
running ISC 2.02, zmdm2.02 and some kind of internal modem, at 2400 baud.

   In both cases, uucp accounts for most of the traffic, zmodem is used
only when file transfer is an afterthought after logging in.

   Perhaps the baud rate is the problem.  Fielding interrupts at
19,200 baud is a tricky business, especially when there are no
required ACK's in the protocol to throttle reception rates down.
While I have no personal experience, everything I've seen in various
newsgroups on the net would lead me to believe that neither VAXen nor
Suns are capable of reliable streamed reception at 19,200 and even
9600 is *really* touchy.

    Of course, in the mentioned case of the VAX 8600, it is getting
LAT packets from the DECSERVER rather than interrupt per byte.  The
machine with the strain is the DECSERVER itself.  I have no knowledge
about its capability to deal with high baudrate incoming data streams.

-- 
# Harvey Moran  moran at tron.UUCP  moran at tron.WEC.COM   ...!uunet!tron!moran
# Westinghouse Electric Corp., Electronic Systems Group, Baltimore, Md.