Print Queue Always Down

[alangoh]

Hi everyone,
I am facing a problem on my IBM AIX machine whereby:-

I install three print servers and each print server is connected to two printers. Each printer is a print queue .

After i run the SMIT to create and install the print queue,everything works fine.But one or two days later, the print queue went down for no reason and print jobs is stuck there.

If i am lucky i need to cancel the print jobs only and everything will works fine. Sometimes i need to delete the print queue and add print queue again in order to make it works again.

Can i know why is this happen and USER IS COMPLAINING!!



FYI, i am still new to IBM AIX.

Cheers
ALAN
NETWORK ALLIANCE TECHNOLOGY

 

[AIXSPadmin]

Most likely the problem is GIGO (Garbage In Garbage Out). Try sending print jobs to the printer other than what the user is sending and see if the same problem occurs. Most likely it will not and then you will have to examine what documents are being printed to isolate the error.

 

[alangoh]

Should i try the solution explained
at
thread52-79377

http://www.tek-tips.com/gviewthread.cfm/pid/52/qid/79377

 

[jpor]

How are the printers connected ?

When I use smit spooler to create printers I always use HP Jet Direct as all of our Printers are on HP JetDirect Boxes/cards.

Once I have added them then I go into /dev and do a mknod <Printer> p to create a device. Then I vi /etc/inittab add the entry to respawn the device. Then run a telinit command to check the entry is correct.

 

[alangoh]

The printer is connected to a three ports print server.
Each printer connected to the print server port is a print queue in the IBM.

 

[jwtesch]

I assume that you are printing with lpr/lpd protocol (backend is rembak or piorlfb) The problem is that when a print server sends a negative acknowledgement NAK) that AIX doesn't retry. With the System V printing in AIX 5L you can specify the time to wait to retry, and the number of times to retry. There is also a program that you may be able to get through IBM services called rembak retry which allows the AIX print queue to be setup the same way.
What is happening is that the buffer on the print server box is getting full from one print job, when another one comes in, it uses NAK to tell the client to wait a minute, but AIX interprets this to mean it is not ready and takes the queue down. The errors are not defined well enough in RFC1179 so neither vendor admits they are wrong.

 

[Yegolev]

basically you will have to restart the queue yourself. i like using 'enq -U -Pqueuename'. implementing a queue monitor is not a bad idea anyway, so you can have it start a down queue if it finds one.

IBM Certified -- AIX 4.3 Obfuscation