Re: netcat delays between pages over wan
From: Bill Vermillion (bv_at_wjv.comREMOVE)
Date: Thu, 25 Mar 2004 16:15:01 GMT
In article <email@example.com>, Gary
Quiring <firstname.lastname@example.org> wrote:
>I am running Netcat on OpenServer 506a. It runs fine locally via
>SCO but over a T1 WAN connection it pauses about 10-30 seconds
>between pages. I use Lantronix EPS2 print servers with HP 5si &
>The file being printed is PCL from a forms tool called Optio. I
>intercepted the PCL file and have a local copy of it. I can print
>the file locally and it's full speed. Send the print over the
>WAN and it has large delays between the pages. Pinging the print
>server over the WAN shows 25ms no packetloss.
All that is going to show you is that ping in the default mode has
not packet loss and that you have a valid connection.
> I tested the same print with three different printers over the
>WAN and they all have the same problem. All of the print servers
>had the same ping results.
And do you have control over this WAN - eg is it point-to-point -
or is it over the 'net - and is it public or in a VPN tunnel mode
through any other devices.
>If I print the PCL file to a Samba printer over the WAN to a
>shared Windows printer it works fine, no delays. All printers,
>servers and Lantronix servers were rebooted.
So you are saying the smb protocol works but data to IP does not?
It might help if you mentioned the print server by make/model.
Now as to the above where I said ping will only show you if you
have connectivity, you need to change your ping options.
By default ping sends 56 byte packets 1 second apart. I had a
similar problem with a client machine getting slow times on the net
and he was saying pings were fine.
So I remoted into the local lan and started checking - and it
turned out his machine - different *i*x OS - responded differently.
All machines same-same and same-diff responded the same to pings
but file transfers via ftp had widely disparate times.
So I started timing via ping and slowly increased the packet size.
The same-same [machines I controlled and new intimately] just kept
up with the load and never started dropping ANY ping packets until
I increased the ping packet size to almost 10k from the default 56.
And then it would drop occasional packets.
But the same-diff [my machine to client] started showing dropped
packets as early as low 250 byte sizes - very few - but some.
By the time I increased packet size to 1000 I was losing well over
60% and it became worse as packet size increased.
So try ping -s [packetsize] with and increase it up to at least the
MTU of the network and see if you still get 100% transmission.
You also might try a flood ping, as even the above still puts in
the 1 second delay and a flood ping will send packets as fast as
possible. USE THAT WITH CAUTION!!
Trouble shooting an ATM link that would just pause for exactly 16
second after ever 1 million byte were transmitted, the telco tech
set up a loop so I could send from an assigned IP address, loop
from Orlando, to Atlanta, to Birmingham, and back to Orlando on a
second IP address on my machine. [it was the only loop available].
He monitored and could see the data start, stop and resume, but
nothing else. Had a Cisco tech go into the intervening router on a
serial port and we did a flood ping of 10,000 byte packets. We
killed the router in under 5 seconds. That was a bug in the IOS that got
I relate that to show you just what havoc you can wreak with a
flood ping - so make sure you use it carefully and let the people
on your net know that it may get slow for a few moments.
Also see other recent threads here about netcat not working will
all of the newer devices. I had a client just change a fairly
large sized Lanier [Ricoh] color laser, and they could not print
from SCO that had been flawless before. They changed the native
protocol in that printer and netcat won't work there.
-- Bill Vermillion - bv @ wjv . com