Bug 398486

Created attachment 114895 [details]
strace -o /tmp/log -f okular --print xanes.pdf

I have gzipped the strace as it was over 4000k

Thanks for your report. A few more questions:

A) Does printing work when you enable "Force rasterization" in the PDF options in the print dialog (this should make Okular use the "default" Qt way of printing).

B) Do other Qt/KDE applications work? And if so, how much works there? (e.g. I'd usually expect that printing itself should mostly to the same extent as with "Force rasterization" in Okular, and the greyed out options are probably always the same?)

c) What Qt version are you using? (There were quite some improvements to the Qt print dialog in Qt 5.11).

I giving the informations you requested below, however I solved the problem and it is a concurrency of problematic server configuration and maybe problematic printing client behaviour.

So. I inherited this server I can admin it but I have no idea how I was set up. I noticed in the server log that I was getting a lot messages like:
E [12/Sep/2018:03:42:14 +0200] Request from "134.157.17.66" using invalid Host: field "rutile1.impmc.jussieu.fr:631"
134.157.189.14 - - [12/Sep/2018:04:31:30 +0200] "POST / HTTP/1.1" 200 418 Create-Printer-Subscription client-error-not-possible
Which made me realize that the printing server is associated with two host names (rutile1 and printers1) on the same IP, but that the cups server was explicitly bound only to printers1. For some reason, even if I've put "printers1" in client.conf, the call to the connection to the server was done by IP, that was interpreted by the server as being on the other hostname (rutile1) and refused.

I've added a ServerAlias with the other host name in cupsd.conf and now everything seems to work. I'm not closing the issue directly, I'd rather let you decide if this is a bug in the printing library or not.

Thank you for you assistance!


1. No: I don't get the error from lpr, but the job does not get in the queue, this is what I get in the strace (I had to use a different printer because the other broke down):
858   setsockopt(19, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
858   setsockopt(19, SOL_SOCKET, SO_REUSEPORT, [1], 4) = 0
858   setsockopt(19, SOL_TCP, TCP_NODELAY, [1], 4) = 0
858   fcntl(19, F_SETFD, FD_CLOEXEC)    = 0
858   fcntl(19, F_GETFL)                = 0x2 (flags O_RDWR)
858   fcntl(19, F_SETFL, O_RDWR|O_NONBLOCK) = 0
858   connect(19, {sa_family=AF_INET, sin_port=htons(631), sin_addr=inet_addr("134.157.17.150")}, 16) = -1 EINPROGRESS (Operation now in progress)
858   fcntl(19, F_SETFL, O_RDWR)        = 0
858   poll([{fd=19, events=POLLIN|POLLOUT}], 1, 250) = 1 ([{fd=19, revents=POLLOUT}])
858   getpeername(19, {sa_family=AF_INET, sin_port=htons(631), sin_addr=inet_addr("134.157.17.150")}, [256->16]) = 0
858   gettimeofday({tv_sec=1536738525, tv_usec=215459}, NULL) = 0
858   getpid()                          = 858
858   openat(AT_FDCWD, "/tmp/0035a5b9c0e80", O_RDWR|O_CREAT|O_EXCL|O_NOFOLLOW, 0600) = 20
858   sendto(19, "GET /printers/leduc.ppd HTTP/1.1"..., 137, 0, NULL, 0) = 137
858   poll([{fd=19, events=POLLIN}], 1, 60000) = 1 ([{fd=19, revents=POLLIN}])
858   recvfrom(19, "HTTP/1.1 400 Bad Request\r\nDate: "..., 2048, 0, NULL, NULL) = 555
858   close(20)                         = 0
858   unlink("/tmp/0035a5b9c0e80")      = 0
858   close(19)    


2. Neither, this was printing from kwrite, the strace is very similar:
socket(AF_INET, SOCK_STREAM, IPPROTO_IP) = 18
setsockopt(18, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
setsockopt(18, SOL_SOCKET, SO_REUSEPORT, [1], 4) = 0
setsockopt(18, SOL_TCP, TCP_NODELAY, [1], 4) = 0
fcntl(18, F_SETFD, FD_CLOEXEC)          = 0
fcntl(18, F_GETFL)                      = 0x2 (flags O_RDWR)
fcntl(18, F_SETFL, O_RDWR|O_NONBLOCK)   = 0
connect(18, {sa_family=AF_INET, sin_port=htons(631), sin_addr=inet_addr("134.157.17.150")}, 16) = -1 EINPROGRESS (Operation now in progress)
fcntl(18, F_SETFL, O_RDWR)              = 0
poll([{fd=18, events=POLLIN|POLLOUT}], 1, 250) = 1 ([{fd=18, revents=POLLOUT}])
getpeername(18, {sa_family=AF_INET, sin_port=htons(631), sin_addr=inet_addr("134.157.17.150")}, [256->16]) = 0
gettimeofday({tv_sec=1536739185, tv_usec=653086}, NULL) = 0
getpid()                                = 1051
openat(AT_FDCWD, "/tmp/0041b5ba2be8f", O_RDWR|O_CREAT|O_EXCL|O_NOFOLLOW, 0600) = 19
sendto(18, "GET /printers/leduc.ppd HTTP/1.1"..., 137, 0, NULL, 0) = 137
poll([{fd=18, events=POLLIN}], 1, 60000) = 1 ([{fd=18, revents=POLLIN}])
recvfrom(18, "HTTP/1.1 400 Bad Request\r\nDate: "..., 2048, 0, NULL, NULL) = 555
close(19)                               = 0
unlink("/tmp/0041b5ba2be8f")            = 0
close(18)    


3. Version 5.11.5
Name        : libQt5Core5
Version     : 5.11.1
Release     : 4.1
Architecture: x86_64
Install Date: Thu 06 Sep 2018 15:42:26 CEST
Group       : Development/Libraries/X11
Size        : 5387516
License     : LGPL-2.1-with-Qt-Company-Qt-exception-1.1 or LGPL-3.0-only
Signature   : RSA/SHA256, Mon 03 Sep 2018 11:31:49 CEST, Key ID b88b2fd43dbdc284
Source RPM  : libqt5-qtbase-5.11.1-4.1.src.rpm
Build Date  : Mon 03 Sep 2018 11:17:12 CEST
Build Host  : lamb05
Relocations : (not relocatable)
Packager    : https://bugs.opensuse.org
Vendor      : openSUSE
URL         : https://www.qt.io

Thanks for the quick update and good to hear you could solve the issue.

(In reply to lp from comment #3)
>
> I've added a ServerAlias with the other host name in cupsd.conf and now
> everything seems to work. I'm not closing the issue directly, I'd rather let
> you decide if this is a bug in the printing library or not.

From what I understand, the underlying problem was mostly a configuration issue, so I'm closing this bug now.

Please just reopen or create a new one should the problem reappear.