Originally posted by: SystemAdmin

I have about 180 printers setup on an AIX 5.3 box. All printers have HP JetDirect cards, so I create them using the command:

/usr/lib/lpd/pio/etc/piomkjetd mkpq_jetdirect -p 'generic' -D ascii -q 'test'
-h 'test' -x '9100'

So I have one queue per printer.

The problem I'm having is that some of the queues are going into the DEV_WAIT state and aren't going back to normal operation.

I've googled and I see recommendations of increasing the timeout value. Problem is, for piojetd backend, there doesn't appear to be a timeout option. Correct me if I'm wrong?

The other "fix" I've found is to do a qadm -K queue; qadm -U queue. That doesn't work either when there is a print job waiting in the queue.

Is there a better approach of printing to 180 printers from AIX 5.3? How can I work around this DEV_WAIT problem?

BTW, sorry about the stupid name. The forum won't let me change it for some reason.....
#AIX-Forum

Originally posted by: SystemAdmin

The DEV_WAIT status code:

"indicates that either the printer is not online, out of paper,has a paper jam, or any similar problem which will prevent the job from printing. Normally, the problem that causes this state has also caused a message to be sent to the job owner or the operator."

The above is a quote from the red book on the subject. I would suggest that there's no point increasing timeouts in this case as you're waiting for something that may not happen until God knows not when. The solution is to look at what is, or isn't, happening at the printer when this comes up. A lot of the causes are a normal part of life that somebody, somewhere, somehow will clear when they need to. A script to "qadm -U" all your printers periodically is a reasonable thing to do. When "qadm -U" works that means the printer is ready, if it fails it isn't.

HTH

Jim Lane
#AIX-Forum

Originally posted by: SystemAdmin

Normally I consider JetDirect queues to be more reliable than LPR/LPD queues in that they don't continuosly go down when there is a timeout and they seem to recover more easily. The code that is run 'piohpnpf' comes from HP and is a pretty simple network socket program. It doesn't really support two way status and all of that though.

Are these really HP JetDirect printers, or are you using this software for other printers? This can often be an issue. Normally you should not see this problem with HP printers as they won't drop the connection. Have you determined if this is happening when the printer is out of paper or ink, or is this related to other unknown issues. It seems strange that this many problems should be occuring with JetDirect printers.

If you have problems getting the queue to go down, sometimes it is easier to kill the piohpnpf application that is sending data to the printer. When a queue is printing, there is a file in the /var/spool/lpd/stat directory that starts with a p.queue.quedev. This contains the PID of the piobe or whatever backend process is controlling that queue. If you kill this PID, it will take down the queue and all the programs running and the rest of the jobs will remain queued. I think the current job will remain queued also.

Some years ago, I wrote a backend that actually restarted rembak jobs when the queue went down, but this may be included today. I haven't checked lately as I don't do a lot of printing anymore.

LAIX Software at the following URL may have a custom solution at a low price that would give you a backend that did an automatic restart. A lot of customers have seemed to have had good success with these solutions.
http://mysite.verizon.net/resobxkm/index/

John Tesch - Redbook author so blame me if things are wrong there.
#AIX-Forum