CUPS 1.2.4 and Windows client -> 1.2.2 me-too

blinkety bill printers at ph.unimelb.edu.au
Tue Oct 31 15:34:47 PST 2006


Well, its interesting, so I'll keep going, please see added info at bottom.

> Please see discussion below (mostly to myself it seems)
>
> (I reordered my previous comment to the bottom below, since I see now that it would get pretty confusing to read if I keep top-posting new info.)
>
>
> >
> >
> >
> > > > > I am running CUPS 1.2.4 on a Debian Linux box with a networked HP Color Laserjet 3600n.
> > > > >
> > > > > I am using hplip the hplip driver (and the recommended ppd).
> > > > >
> > > > > No matter what I try I simply cannot get a Windows XP client to successfully install the native driver and use this printer.
> > > > >
> > > > > Each time I try to install the driver, the Windows printer install wizard hangs as soon as I select the 3600n driver. I have to kill the spooler to recover things.
> > > > >
> > > > > I have looked at my permissions in my cupsd.conf, and they look fine. I have also made sure mime.conv and mime.types have the octet-stream lines uncommented.
> > > > >
> > > > > I've even tried install a generic postscript driver on the windows client, but it also hangs.
> > > > >
> > > > > I've never had trouble with CUPS and native drivers on windows clients before, but this is the first time I've used 1.2.4 and also the first time I've used hplip driving a network printer.
> > > > >
> > > > > Advice appreciated.
> > > > >
> > > > > Cheers, Iain
> > > >
> > > >
> > > > Sorry, no advice, except I doubt it's your fault.
> > > >
> > > > I have the same problem, with CUPS 1.2.2 on Ubuntu. I see the exact same thing, with HP 2420 and HP 9050, both of which were working last week ...
> > > >
> > > > But now I've reinstalled my test client, and it seems to me something is wrong on the server. Possibly something over the weekend, a silent upgrade or crash.
> > > >
> > > > The server has had some network probs, and I wonder if something is clogged as a result. I've had the service restarted. Also the clock has been skewed out by a day or so, could that cause a problem?
> > > >
> > > > Network seems OK now, because other printers can print, and Linux users can print to one of the 9050 printer at least (haven't tested the 2420's).
> > > >
> > > > On the 2040, I've tried using HP PS and PCL6 drivers; On the 9050 I've tried using PS, PCL, Xerox 4400N, cups test v6 drivers.
> > > >
> > > > Its very odd, because I CAN print OK and get properties page etc for several other printers including Xerox 4400N PS and HP 2200D PCL.
> > > >
> > > > So it doesn't seem to be that the HP driver on the linux server is the problem .. unless it's convoluted enough to have printer-specific problems that is.
> > > >
> > > > Well, I'm sure its not that, because if I create a RAW queue for the 9050, it doesn't work any better for the windows client.
> > > >
> > > > I've been assured that nothing has changed on the server, but it's another admin who look after it.
> > > >
> > > > I have tried various of the windows installations on several client machines, each exactly the same. I get through the install if I use a RUNDLL command-line in a script to add the printer. But then, accessing the properties page doesn't succeed: restart spooler, see error message: "Printer properties cannot be displayed. Operation could not be completed."
> > > >
> > > > If I install the driver pointing to a bogus server, the message comes back immediately, instead of the indefinite wait until I restart the spooler.
> > > >
> > > > If I point to a real server, but a bogus printer name, I get an error during installation .. which at least is one thing that does make sense.
> > > >
> > > > Any suggestions received gratefully.
> > > >
> > > > Blinkety Bill
> > >
> > > I should have mentioned, I am installing the printers as IPP ('network' http printers) .. and as I mentioned this was working in testing, and 2200 and xerox printers love it.
> > >
> > > BB
> >
>
>
>
> > More info:
> >
> > I have an image of a system that was configured last week when 'all was well'. This system *works fine* with the HP printer in question.
> >
> > *UNTIL*: If I add a new instance of the printer -- using *exactly* the driver which is already installed, not even providing the files again, but choosing the existing driver in the list of known system printer drivers when offered in the windows GUI -- it hangs!
> >
> > After I stop and restart the spooler, both instances of the HP printer are now broken - the spooler hits 97% CPU and stays there until restarted yet again. But other printers continue to work, as described in my previous post (see way below).
> >
> > So there is something which is preventing the driver from resolving *something* to do with the printer that it must do the first time. Whereas, the driver is quite happy to work if already installed.
> >
> > I know there have been no new windows updates *at all* issued, the client PC image is pristine and exactly identical to the config it had when I installed the printer onto it last week.
> >
> > Does anyone have any idea what would cause this?
> >
> > Blinkety Bill
>
> And now one of the Xerox printers is doing this. But only one of them .. in fact only one class queue.
>
> In this case, if I delete and recreate the class, it does this still. If I create the class with a different name, it does not do it.
>
> I am also deleting and reinstalling the printer in windows each time. The same thing happens if I create a duplicate of the printer in windows, so I guess its not some gumming up of (settings <=> name) linkage ...
>
> This is slightly different to with the HP printer. In that case, the printer queue itself is gummed up. The class is gummed too, but I guess that can't be seperated from the printer being gummed.
>
> We have shut down cupsys and restarted it. No zombie threads seem to be running. We also killed and restart the hplips service, I think that's used by CUPS. But now that a Xerox is also gumming ...
>
> However there does seem to be *some* level of specificity to HP printers, ie I installed a number of 2420s as well as one 9050, and *all* of those gummed this week (but not last week!).
>
> Blinkety Bill

OK, the Xerox class is only a problem from one specific windows machine. I installed it on another box, and it works.

Whereas the HP 9050 is a problem from all of the windows clients I've tried it with as described above -- ie it works in the one that was already installed, but not in any, including that one, if newly installed.

OK, enough. I will probably get a job as a waiter.

Blinkety Bill




More information about the cups mailing list