QUEUE paused: Unable to send print file to printer: Connectionreset by peer

[Helge Blischke]

"Vanush (Misha) Paturyan" wrote:
> 
> Helge Blischke wrote:
> > Gennady wrote:
> >
> >>Sometimes (3-4 in day) printer queue paused, when I start printer - all works fine, spooled documents printed normaly.
> >>
> >>In log I found this message:
> >>
> >>Unable to send print file to printer: Connection reset by peer
> >>
> >>Printer:
> >>Device URI: socket://192.168.0.5:9100
> >>
> >>Search for cups.general without success.
> >>What can I do in this situation?
> >
> >
> > Buy a new printer.
> 
> This is lamest suggestion I've ever heard!! Are you telling me I  have
> to replace 5 printers in here because CUPS cannot properly communicate
> with them? They all worked perfectly fine with Windows NT 4.0 Server.
> 
> To the point: we have mixture of HP LaserJet 4050, 4100 and 4250 in
> here, all do that sometimes. Mostly then someone sends a 10Mb pdf file
> to the printer, or when printer runs out of paper and nobody replenish
> it. I suspect there should be some settings to tweak timeouts. I looks
> like CUPS waits for printer to respond, printer is busy doing something,
> and then printer actualy reponds CUPS already put printer into
> "disabled" mode and won't even talk to it.
> 
> What exactly Timeout parameter in the cups.conf do?
> 
> > Well, no kidding. The message you get occasionally tells you that the printer closes the connection
> > when it should not do so. CUPS cannot but treat this as an error and stops the queue to prevent
> > (further) losss of data to be printed.
> >
> > Helge
> >
> > PS: to be honest, the error may be caused by a bogus router or weak hardware somewhere in your
> > network, of course.
> >

Another thought on this issue: the behaviour may depend on the revision of the JetDirect firmware
your printers use. Perhaps you could try the hpnpf backend instead fo the socket backend - hpnpf
issues a "@PJL USRSTATUS JOB=ON", which forces, if interpreted correctly, the printer (the JetDirect
interface, to be precise) to report the status WHEN THE JOB HAS FINISHED to the requestor and thus
MUST hold the connection open until the job is done.

Our experience at our site shows that the only printers that occasionally show the bogus behaviour
you described are the (very few in the meantime) old HP4M and HP5M printers, which reportedly do
not honour those status commands.

Helge

-- 
Helge Blischke
Softwareentwicklung
SRZ Berlin | Firmengruppe besscom
http://www.srz.de