lpsched

[Shaamil]

Hi All
Does anyone know why lpsched would hogg my system. 96% cpu usage.

I am running a Service Gaurd cluster on a 2 node K570 servers running 10.20.
When this happens, the system is dead slow. I have to kill all the lp processes and try to startup lpsched again.

Thanks in advance
Shaamil

[Uday_S_Ankolekar · Accepted solution]

Most of the time this is becasue of wrong printer name , it has to be the same as the name in /etc/hosts or in DNS associated with the ip address of your printer.
This could happen if an incorrect peripheral name for your printer,you must remove and readd the printer to the spooler using the correct peripheral name .
Goodluck,
-USA..

[Shaamil]

I will check with my application users, which printer it is that they are using. We have about 20 printers on the system.
Thanks,
Shaamil

[Robert Salter]

Did you run glance or top to see which lp process was taking up the CPU? Or if it was more than one process. Could be a application problem constantly sending print jobs. What does 'lpstat -t' show?

[Bill Hassell]

The lpsched consuming all the CPU time will be the parent of a shell that is running the shell script.The shell script is the name of the local printer. To see the hierarchy of processes, use this:

UNIX95= ps -ulp -H -o pid,ppid,time,args

The highest level lpsched has a parent ID (PPID) of 1 and all other lpscheds will have this first lpsched as their parent. Look for the lpsched that is taking all the CPU time, then look at it's child process which will be the shell running the printer script. That printer is the problem. Fixing it depends on how it is connected, local via copper cable, local via HP JetDirect networking or remote to a non-HP printer or remote print server.

[Shaamil]

Thanks for the response guys. My printers are all idle for the weekend, but I will let U know how it goes next week. I am away next week as well.