Raster filter fails when printing on shared printer from network

Bug #1811626 reported by Ovidiu-Florin BOGDAN
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
HPLIP
New
Undecided
Unassigned

Bug Description

My set-up:
Server machine:
    * Ubuntu 16.04.5
    * CUPS 2.1.3
    * HPLIP 3.16.3
    * an HP LaserJet 1018 connected via USB and shared in the network via CUPS
Printing works just fine directly form the server machine (using lp command).

Client machine:
    * Fedora 29
    * CUPS 2.2.8
    * HPLIP 3.18.6
    * the shared printer is installed here via connection: dnssd://HP%20LaserJet%201018%20%40%20ServerName._ipp._tcp.local/cups?uuid=XXXX
Printing from the client results in the job being stuck on the server with the following log snippet:

D [14/Jan/2019:00:05:01 +0200] [Job 20] Started filter /usr/lib/cups/filter/hpcups (PID 14897)
D [14/Jan/2019:00:05:01 +0200] [Job 20] Started backend /usr/lib/cups/backend/hp (PID 14898)
D [14/Jan/2019:00:05:01 +0200] [Job 20] prnt/hpcups/HPCupsFilter.cpp 565: cupsRasterOpen failed, fd = 7
D [14/Jan/2019:00:05:01 +0200] [Job 20] prnt/backend/hp.c 919: ERROR: null print job total=0
D [14/Jan/2019:00:05:01 +0200] [Job 20] PID 14897 (/usr/lib/cups/filter/hpcups) stopped with status 1.
D [14/Jan/2019:00:05:01 +0200] [Job 20] Hint: Try setting the LogLevel to "debug" to find out more.
D [14/Jan/2019:00:05:01 +0200] [Job 20] PID 14898 (/usr/lib/cups/backend/hp) exited with no errors.

This bug report comment [1] advises not to install a driver on the client machine, which I find odd, because then I cannot control how the print job is to be printed, since CUPS doesn't know anything about the printer.

I'll look more into this and keep this bug report updated, but I could sure use some help, since the HPLIP resources are scattered and there is no clear central point of management for all this.

Since I see that `cupsRasterOpen` is defined in the CUPS sources, I guess I need to look there as well, since the error for the file descriptor open() call to be printed, so I guess the file in question exists just fine. Now that I think of it, this might actually be a CUPS bug.

[1] https://bugzilla.redhat.com/show_bug.cgi?id=1010580#c9

Revision history for this message
Ovidiu-Florin BOGDAN (ovidiu-florin) wrote :

On the local client I have the following log snippet when I print a test page:

ian 14 01:25:41 Galago cupsd[2497]: REQUEST localhost - - "POST /printers/HP_LaserJet_1018_Apollo HTTP/1.1" 200 430 Print-Job successful-ok
ian 14 01:25:41 Galago hpcups[21139]: common/utils.c 173: validate_plugin_version() Plugin version[3.18.12] mismatch with HPLIP version[3.18.6]
ian 14 01:25:41 Galago hpcups[21139]: common/utils.c 206: Plugin version is not matching
ian 14 01:25:41 Galago hpcups[21139]: prnt/hpcups/HPCupsFilter.cpp 489: m_Job initialization failed with error = 48
ian 14 01:25:41 Galago hpcups[21139]: common/utils.c 277: Invalid Library hanlder pLibHandler = NULL.

I've tried running `hpsetup -u -i` and selected the network connection type, but it fails to detect anything and exits with: error: No device selected/specified or that supports this functionality.

description: updated
Revision history for this message
brian_p (claremont102) wrote :

I would advise that you remove HPLIP and the plugin software from the client, read

https://wiki.debian.org/DriverlessPrinting

and set up an everywhere queue with your present dnssd://... URI (or an ipp://... one) or allow cups-browsed to do it for you. This system will produce a PDF to send to the server. The server will happily process it.

> This bug report comment [1] advises not to install a driver
> on the client machine, which I find odd, because then I cannot
> control how the print job is to be printed, since CUPS doesn't
> know anything about the printer.

Perhaps

https://wiki.debian.org/PrintQueuesCUPS#Double_Filtering

explains why it is not advised.

--
Brian.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.