The bug is rock solid reproducible and has been around for about 7-10 days.

With a cold boot and nothing in the CUPS print queues; (lp1 double sided A4, lp2 single sided A4) when I "print" a document it gets no further than the print queue. and there is no /dev/lp0. same if I do a test print from within CUPS is queued. Restarting the spooler from inside the CUPS GUI had no effect.

If I then reboot the m/c. I can now see /dev/lp0 (parallel port) is now present,
crw-rw---- 1 root lp 6, 0 2010-02-01 14:21 /dev/lp0,
the documents in the queue from before the reboot are now printed,
and I am able to print any further documents until the next reboot.


The kit,:
K7S6A with 1GB of RAM and 4 IDE disks.
o/s is : “Linux tigger 2.6.26-2-686 #1 SMP Sun Jun 21 04:57:38 UTC 2009 i686 GNU/Linux”
print chain: CUPS and HP 4050TN connected to parallel port by cable (D25 to Centronics)



Plenty of disk space
df -k
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/hda3 385323276 5274280 380048996 2% /
tmpfs 518044 0 518044 0% /lib/init/rw
udev 10240 100 10140 1% /dev
tmpfs 518044 0 518044 0% /dev/shm
/dev/hda1 459143 6549 428097 2% /boot
/dev/hdb1 390696872 7083516 383613356 2% /home/spirals
/dev/hdc1 312559096 83760504 228798592 27% /home/spirals/Desktop/work
/dev/hdd1 390696872 382188 390314684 1% /usr/src




My work round is to “print” a file with printer off before shutdown, a waste of paper but a temporary fix.

Found following error messages: in /var/log/cup/ [I was trying to print something off the web, it locked I killed it with SIG9].

error_log.4
E [26/Jan/2010:12:07:16 +0000] PID 3821 (/usr/lib/cups/filter/pstops) crashed on signal 9!
E [26/Jan/2010:12:07:16 +0000] PID 3822 (/usr/lib/cups/backend/parallel) crashed on signal 9!

error_log.5
E [20/Jan/2010:13:40:17 +0000] PID 3789 (/usr/lib/cups/backend/socket) crashed on signal 9!

error_log.6
E [19/Jan/2010:07:36:42 +0000] PID 8873 (/usr/lib/cups/backend/socket) crashed on signal 9!

Any suggestions as to what I should check or how to fix please.

AFAIK, signal 9 is "segmentation fault", usually faulty RAM. Try checking your RAM

Not very sure, because if the RAM is really faulty, surely you'll get your system hung or something worst than a little printing problem. But, worth to check, IMHO :-)

1) SIG 9 is SIGKILL what I used to kill (sledgehammer) hung processes (as it cannot be ignored), nasty but works every time.

I ran a one hour batter test on the RAM, with no errors, so it looks OK.

I can see the printer being recognized correctly at paraport0 during the boot process.

From the Web interface, trying to print a test page, with a cold booted m/c and emptym print queues, I see the message ""Printer not connected; will retry in 30 seconds..." but the printer is connected and if I then reboot the test page is printed and this message not seen. As already stated, when I then try a new test page it prints and anything else I send to the printer. [printing is normal ?]

In the fault mode looking at the /var/log/cups files:
error.log is empty

page.log reads:
lp1 anonymous 1701 [03/Feb/2010:10:59:18 +0000] 1 1 - localhost

access_log reads:
localhost - spirals [03/Feb/2010:10:56:23 +0000] "GET /jobs/ HTTP/1.1" 200 0 - -
localhost - - [03/Feb/2010:10:56:23 +0000] "POST / HTTP/1.1" 200 438 Get-Jobs successful-ok
localhost - spirals [03/Feb/2010:10:56:23 +0000] "GET /jobs/ HTTP/1.1" 200 3865 - -
localhost - - [03/Feb/2010:10:56:26 +0000] "GET /printers/ HTTP/1.1" 200 0 - -
localhost - - [03/Feb/2010:10:56:26 +0000] "GET /printers/ HTTP/1.1" 200 15767 - -
localhost - - [03/Feb/2010:10:59:17 +0000] "GET /printers/lp1?op=print-test-page HTTP/1.1" 200 0 - -
localhost - - [03/Feb/2010:10:59:17 +0000] "POST /printers/lp1 HTTP/1.1" 200 17551 Print-Job successful-ok
localhost - - [03/Feb/2010:10:59:17 +0000] "GET /printers/lp1?op=print-test-page HTTP/1.1" 200 3397 - -
localhost - - [03/Feb/2010:10:59:20 +0000] "GET /printers/lp1 HTTP/1.1" 200 0 - -
localhost - - [03/Feb/2010:10:59:20 +0000] "POST / HTTP/1.1" 200 454 Get-Jobs successful-ok
localhost - - [03/Feb/2010:10:59:20 +0000] "GET /printers/lp1 HTTP/1.1" 200 7955 - -
localhost - - [03/Feb/2010:10:59:25 +0000] "GET /printers/ HTTP/1.1" 200 0 - -
localhost - - [03/Feb/2010:10:59:25 +0000] "GET /printers/ HTTP/1.1" 200 15860 - -
I presume they are OK?

Anyone able to say quickly which pins on the D25 and Centronics plugs do I test continuity through, in case I have a cable fault? I do not believe I have as the reboot clears the fault. I suspect it is software but (?) What else can I try? any pointers appreciated. And thank you all.

(head scratch and frown - no icon available)

While searching line by line in syslog found with printer off "cwdeamon" was trying to seize the paraport0 with exclusive access at boot time. Why it was doing this is unknown but deleted cwdeamon package and all is back to normal. Thank you all for help.regards, (spirals -dja.)