spooler won't start

[HU YIHONG]

My system is a 9000/777 workstation running HP-UX 10.20. One day we had some problems with printing so I stopped spooler by SAM. When I tried to restart spooler I failed. I even tried rebooting the system. I noticed that a SCHEDLOCK and core were generated under /var/spool/lp/ whenever I run lpsched.
I tried to remove SCHEDLOCK before start the scheduler with lpsched but SCHEDLOCK would be created again together with core.
Please help and thanks.

[KCS_1]

Hi,

Under the url is explain about SCHEDLOCK and FIFO.
I hope the url is better help to you.

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x30f6f841489fd4118fef0090279cd0f9,00.html

have a nice day!!

[HU YIHONG]

Dear Patrick

Thank you for your help.
The explanation helps me understanding better of SCHEDLOCK and FIFO. However when I tried lpsched it said scheduler is running. Then I couldn't find lpsched by ps command. If I run lpshut it would complain no scheduler running.

[Steve Steel]

Hi

1)Cleanup
# /usr/sbin/lpshut -> shut down the scheduler
# ps -ef | grep lp -> are lp processes still running ?
-> if yes, kill those processes
# kill -> kill ALL lp processes
# ps -ef | grep lp -> check if all processes are really gone
# cd /var/spool/lp
# rm FIFO SCHEDLOCK CLD_FIFO Toutputq *.lock
-> delete temporary files, not all of them may exists

2)Restart
# /usr/sbin/lpsched -v -a -> restart lpsched with additional logging enabled

3)Check
Look in /var/adm/lp/log for info
If corefile do file core and post output
What is the last lp patch on the box

swlist -l fileset|grep -i lp


Steve Steel

[Ulf Lipski]

Hi,

I suggest the problem is that you have corrupted *status files. So remove them and restart the scheduler. If necessary to do this remove first the *FIFO and *SCDEDLOCK files.

By

[HU YIHONG]

Hi Steve and Ulf

Thank you for your help. I tried but couldn't get anywhere. I guess some files got messed up. I used SAM to stop the spooler and couldn't start it since then. no messages logged under /var/adm/lp/log after that.

lpsched creates a core dump every start.
I attached output from strings core here.

[Scott_14]

Hello:

I may be off base with this, but when I was at one place on 10.2, with a lot of queus, this would happen to me. After removing the lock file, and fifo, I would run ps -ef |grep lp
and see other scheds running, i would kill all of them, and run a cancel on all queus, this seemed to work, why I am not sure, but it would seem after canceling the queues I could get the lpsched to run again.

Not sure if you want to try that, but if your lpsched is still down, it may be worth the try.

scott

[Steve Steel]

Hi

1)Please file core

Note strings core.

2)lpstat -t does it work.

3)Nothing in the log suggests a protection error that lp cannot write.

I have uploaded a script

It correct permissions.


Steve Steel

[Steve Steel]

Hi

Another script

#!/bin/ksh
#
#Printer parameter
#
# parameter is enable disable accept or reject to do all printers
#
option=$1
xx=`lpstat -p|grep printer|cut -f2 -d" "`
for printer in $xx
do
case $option in
accept) accept $printer ;;
reject) reject $printer ;;
enable) enable $printer ;;
disable) disable $printer ;;
*)echo invalid parameter ;;
esac
done


Save and use as script disable

Will disable all printers

Then try and start the spooler

When running script enable to restart printers

Steve Steel

[Frank Slootweg · Accepted solution]

If these printers were managed by SAM, i.e. *not* by lpadmin commands, or at least were managed *last* with SAM (*and* correctly working at the time), then I advise to use SAM's Restore Spooler Configuration state functionality:

sam -> Printers and Plotters -> LP Spooler -> Save/Restore Spooler Configuration

Check that the displayed "Time Saved:" is recent/correct!

The actual Restore Spooler Configuration action is under "Actions".

[Victor_5]

Your name let me recall one of my previous co-workers, did you work for CCC of CHP before in HangHua?

[Eric Buckner]

I had this happen a few times to us on 10.20. Make sure there are not schedulers running. And rename /var/spool/lp/outputq to /var/spool/lp/outputq.old. Then start the scheduler. If it doesn't work you can move the outputq file back. But in each time our similar issue was the result of a corrupted outputq.

HtH,
Eric

[Jose Mosquera]

Hi,

You must guarantee that every printer showed in your SAM spool have a corresponding directory in /var/spool/lp/request. If not, pls follow this steps:

mkdir
chown lp:lp
chmod 766

remove SCHEDLOCK file

restart spool

Other way to restore any previous working-fine spool configuration is:
Make periodically in SAM/Printers and Ploters/LP Spooler/Save-RestoreSpooler Configuration/Action/Save Spooler Configuration

When you need restore this configuration just select Restore Spooler Configuration in same path showed above.

Regards,

Jose

[HU YIHONG]

Hi,

It's the spooler configuration. I restore the configuration using SAM and it works.

Thanks to everyone who bother to help me.
I am really touched by the response I get.
It's the first time I seek help here.

Victor, I am not the one you knew.

Have a nice weekend!