Originally posted by: SystemAdmin

Afternoon, all.

We currently run a number of AIX 5 systems serving users at a number of remote sites and are having problems with our AIX print queues.

The queues are all identical (created using a scripted interface to allow the Service Desk to service new queue requests) and look like this;

asystem/usr/home/centre/tech $ lsque -q 0020pr01
0020pr01:
device = @0020pr01
up = TRUE
host = 0020pr01
s_statfilter = /usr/lib/lpd/bsdshort
l_statfilter = /usr/lib/lpd/bsdlong
rq = lp
marge/usr/home/centre/tech $ lsquedev -q 0020pr01 -d@0020pr01
@0020pr01:
backend = /usr/lib/lpd/rembak -T7200

Here '0020pr01' is the hostname for an OKI ML5520 printer - again they are all identical.

The problem we are having is the print queues going down. They are easy enough to start up again but this is a great inconvenience to users and generates a lot of calls to the Service Desk.

Unfortunately I do not fully grasp what would cause these print queues to go down. As far as I understand it there are two types of print queue; 'local' and 'remote', the above belonging to the latter group. Apparently with 'local' type print queues you can set the option 'recovery_type' for the queue, and use this to specify that in the event of the queue going down it tries to start up again every X minutes a set number of times. So, my questions are - what could be causing the queues to go down? And is there any way to set our printers up as 'local' type queues so I can take advantage of the 'recovery_type' option?

Any other pointers would be very much appreciated, pointers to documentation or anything to aid my understanding of the processes involved and how they communicate with one another.

Thanks and regards,

Alex

Originally posted by: SystemAdmin

Alex: a first step would be to generate some debug output. You can specify a log file for rembak in /etc/qconfig using the -D flag. Similarly the lpd daemon can produce a log if you start it as

startsrc -s lpd -a "-D /tmp/lpg.log"

HTH

Jim Lane

Originally posted by: SystemAdmin

Because the RFC does not specifically tell what to do when a printer server wants to wait, some print servers and network printers send a NAK to say their buffer is full. Technically this says they refuse the print job, and AIX then responds to this by taking the queue down. Check with support line on whether you can use the local queue approach with a remote queue with local formatting. Several years ago I wrote a custom backend that would retry a specified number of times and resent the job then save the job and send an email, but I'm not sure if I've still got it.

Originally posted by: SystemAdmin

I have this problem with many of my remote printers that are attached to remote serial print servers. HP Jet Direct and Intel print servers do not have this issue in my network. I run the following script every 10 minutes via cron:

#!/usr/bin/ksh

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