New SNMP printer discovery backend

Kurt Pfeifle kpfeifle at danka.de
Tue Apr 18 17:03:28 PDT 2006


> Kurt Pfeifle wrote:
> > I'm very pleased to have seen the new SNMP printer discovery
> > backend added to 1.2 SVN a few hours ago  :-)
> >
> > My first test result is very encouraging. I compiled (building
> > in the source directory), yet didn't install, but instead did
> > run the backend directly from the build directory like that:
> >
> >   kurt at sp:~/cupssrc> LD_LIBRARY_PATH=$PWD/cups:$PWD/filter backend/snmp @LOCAL
> >
> > Guess what? It works :-)
>
> Yay! :)
>
> > It discovers 10 printers within ~2 seconds:
>
> Right, it will only spend a maximum of 3 seconds to find printers
> via SNMP, and then probes the standard ports (631, 9100, then 515)
> to discover the right URI to use.  As Ira would probably say, we
> should be using the new Port Monitor MIB for this, but currently
> none of the 76 printers I have in my lab supports it... :(

I'll need to find out if any of the devices here do support it
already. If so, I may get the permission to create an SSH tunnel
for you to access one of our subnets here if you think this could
help you with testing. Otherwise, I'd be glad to act as one of
your proxy testers if that is feasible.

> >   kurt at sp:~/cupssrc> time LD_LIBRARY_PATH=$PWD/cups:$PWD/filter backend/snmp @LOCAL|wc
> >   INFO: Using default SNMP Community public
> >   ERROR: Bad SNMP packet from 10.162.2.91!
> >      10      81     994
>
> You can set the CUPS_DEBUG_LEVEL environment variable to "3" to get
> really detailed debug output,

OK.

> plus pass an IP address (or hostname)
> on the command-line to probe a single printer.

I knew that, but I didn't know any of the single addresses yet by
heart (since we moved to a different floor within the building, with
different IP addresses)  :-)

> If you can, please post results for any printers that return an
> error (the "Bad SNMP packet from ..." messages) -

I will.

> I'll update the
> backend to automatically log the packets with errors.

Cheers,
Kurt




More information about the cups-devel mailing list