Message from queueing system

[fredericrenaud]

Hi,

I'm a new AIX admin and I'm having a hard time figuring out how to solve a printer problem.

Our print queues goes down from time to time and we have no idea why. All our print queues are define like this:

/etc/qconfig

mars:
device = @toto
host = toto
rq = mars
s_statfilter = /usr/lib/lpd/bsdshort
l_statfilter = /usr/lib/lpd/bsdlong
@toto:
header = never
trailer = never
access = both
backend = /usr/lib/lpd/pio/etc/piorlfb -f !

toto is a valid hostname and it works fine most of the time and it's never the same queue that goes down.

Our users receives a message similar to this when a queue goes down:

Message from root on saturn (UNKNOWN) [ven oct 21 10:42:41 2005] ...
Message from queueing system:
0781-088 Queue mars septoto went down, job is still queued:
Backend: /usr/lib/lpd/pio/etc/piorlfb -f ! -S toto -P mars sep \
-N /usr/lib/lpd/bsdshort /var/spool/qdaemon/tw2yvia
Backend Exit Value: EXITFATAL (0100)
Job Submit Time: Oct 18 06:00:57 2005. Current Time: Oct 21 10:42:41 2005 Use local problem reporting procedures.
<EOT>

Message from frenaud on saturn (UNKNOWN) [Fri Oct 21 10:47:06 2005] ...
Message from qdaemon:

: (FATAL ERROR) Failed receiving acknowledgement.
<EOT>

Message from frenaud on saturn (UNKNOWN) [Fri Oct 21 10:47:06 2005] ...
Message from qdaemon:
/usr/lib/lpd/rembak -S toto -P mars -N /usr/lib/lpd/bsdshort -o -f -o l
/var/tmp/piorlfb_49024 <EOT>

Message from root on saturn (UNKNOWN) [ven oct 21 10:47:06 2005] ...
Message from queueing system:
0781-088 Queue marstoto went down, job is still queued:
Backend: /usr/lib/lpd/pio/etc/piorlfb -f ! -S toto -P mars \
-N /usr/lib/lpd/bsdshort /var/spool/qdaemon/tr0oUMa
Backend Exit Value: EXITFATAL (0100)
Job Submit Time: Oct 21 10:46:04 2005. Current Time: Oct 21 10:47:06 2005 Use local problem reporting procedures.

I have 2 questions.

Question 1: Any suggestion on how to debug this?

Question 2: Is it possible to capture these messages so that the users won't see them? If so, how?

thanks

ps: a similar thread was posted 2 years ago on the forum, but there were no answers.

Frédéric Renaud

 

[needcoffee]

This problem occurs when the queue goes down usually due to a buffer full problem with the destination printer or print server and the queue is not restarted before the timeout value elapses. The I use the following, found on Tek-Tips but can't remember who to credit, as a root cron job that runs every 5 minutes.

enq -As | grep "DOWN' | cut -f 1 -d " " | while read ln
do
 enable $ln
done

Users will now only see the message when a persistent problem is occuring with the destination deivces such as paper out, paper jam, or loss of power during a job.

For question 2 I think you could redirect stderr to null for the print queue system and the users may not receive the message but I have not tested it. You also would not have any log or notification of errors accross the board for queue issues.

Hope this helps.
needcoffee

 

[fredericrenaud]

We have a similar job (below), but using "enq -As | grep DOWN" is better than "lsallq".

/usr/bin/lsallq|/usr/bin/xargs -I{} nice -n 15 /usr/bin/qadm -U{}

Is there a way to change the timeout setting (to prevent going down)?

about question 2, where could I specify the redirection or the print queue system? in smitty?

Thanks for your help and fast answer.

Frédéric Renaud

 

[needcoffee]

For question 1 change the following line in qconfig:

backend = /usr/lib/lpd/pio/etc/piorlfb -f !

to

backend = /usr/lib/lpd/pio/etc/piorlfb -f ! -T999

This should make the time out value unlimited.

As for question 2 you would either have to start qdaemon or piobe with startsrc and point stderr to /dev/null. I have not tried this on our system nor do I recommend it.
Good luck.

needcoffee

 

[fredericrenaud]

Thanks again for your help.

Frédéric Renaud