lpd problem over wan

[BRAGEUL]

Hello,

I've got this probleme since two week about my 27 lpd printers (Jetdirect Box) Wan connected.

When i launched a printing request on a remote printer, i'v got sometimes this message.

02-Nov-2004 14:28:08 [400] lpd[128715]: ERROR -- read_all_chars: read() failed,
error: Connection reset by peer

And the printer print a truncate page. It's a PCL format document (2 or 3 pages). Sometimes it's look good but the last day it really the panic.

Is there any parameter to tune TCPIP in my Unix server ?

I've got 2 servers DS20E & DS10 with Tru64 5.1A PK6 and 5.1.B PK4.

Thank

[Michael Schulte zur Sur · Accepted solution]

Hi,

does this occur on both servers? I have seen something with this after an installation of pk6 on 5.1A. Can you post the printcap section of one of those printers?

greetings,

Michael

[BRAGEUL]

Thank for your response,

Effectively, i've got this problem with my 2 servers.


Here is my attachment printcap file:


Didier.

[Matt Palmer_2]

Hi,

I've had problems with a Jet Direct 500X over WAN. they were similar errors, also including vanishing print queues. I solved this in the end by reducing the buffer size on the jet direct, so that it had to make more frequent calls back to lpd on the server(this seems to 'hold' the connection open better), otherwise I was suffering connection breaks, as you are.

hope this helps.

regards

Matt

[BRAGEUL]

Hello,

Well ..
I had a look in the web interface of my Jetdirect 170x .

Where i can find this buffer value setting . Is there a specific value of the 500X's model ?

I found a TIMEOUT VALUE : default 1500 seconds . I don't think that it is this paramater ..


Didier .

[Matt Palmer_2]

Hi,

apologies, I've just spoken with my colleague whom I worked with on this, and it was the buffer size on the physical printers which we reduced - not the 500X, and this stopped the connections dropping through the jetdirect.

You are correct, I would imagine that increasing the timeout would only 'mask' the problem, and reducing it certainly exacerbate it.

Is it possible for you to try the buffer reduction on one of the printers??

regards

Matt

[Alexey Borchev]

Looks like WAN problem to me...
I wolud try to monitor network behaviour.
At least by pinging Jetdirect (with packetsize=1k) and seeing responce time.
Or better see router's stats.

[BRAGEUL]

I make a test ping :

Ping -s 1024 192.170.0.100 (Jetdirect Ip )

I 've got a response of 169 ms constant.

Without -s argument : 72 ms

Wich value can i monitor in netstat to see if i've got a network Wan Problem ?



Didier

[BRAGEUL]

I've got this problem with my 27 remote LPD printer

So i make the last day fulle test and here is my result:


I Stop the queue of one of my remote printer to verifyed the file under /usr/spool/lpd1 (for my lp1 printer):

I launch 2 printing job and i had a look to the spool job under this directory.

So : When the print file size job is less than 8ko, the printing process failed !!
when the file size is more than this critical value, the printing is good .

When the file is lower than 8ko, it is a request for printing of a table which covers less than 75% of the page.

Is there any parameter in the printcap file to flush the printing job and to tell to jetdirect box to finish the print process when the file is really little.

Or something else like that .

Any idea .... ?

[Michael Schulte zur Sur]

Hi,

you use the bsd spooling. Have a look into the tcp/ip printing protocol.

greetings,

Michael