lost all output??

[Kylie Spooner]

Hi everyone

I have the good fortune of working on a HPUX 11.0 server that has seemingly died on it's bottom. In my message browser I saw many messages telling me that a process couldn't fork because there are too many processes already running. I did a ps -ef |wc -l and it came back with 693. I closed opc on several of the machines here (hoping to lighten it's load), but it only got gradually worse. Now I am at the point that I got kicked out of the shell and cannot even get any output to the screen. I even went and hooked hyperterminal up to the console port without so much as a prompt appearing on the screen. Does anyone have a clue what might be going on here? Maybe a way to reboot my box without actually hitting the button? I am just a beginner, so sorry if this is an improper question.

Thanks for help in advance,
Melb

[Sridhar Bhaskarla]

Melb,

I congratulate on your fortune. It's really fun to work with HP-UX.

You are running out of nproc kernel parameter value.

Increase this value to 30% of what you have now..

Use "sam" to do this for you. Keep a track of this parameter using sar -v 2 10 looking at
proc-sz.

-Sri

[Patrick Wallek · Accepted solution]

It sounds like you have hit the limit defined by your nproc kernel parameter.

The nproc kernel parameter controls how many processes can be run on the system at the same time.

The way to solve this problem in the short term is to kill some of the processes running on your system by stopping some applications or logging some users off or something similar.

In the long term it sounds like you should increase the nproc kernel parameter. This can be done via SAM->Kernel Configuration -> Configurable Parameters. Here you can set the parameters and SAM will rebuild the kernel for you and reboot your system.

The nproc by default uses maxusers to help figure its value. So to increase nproc, you can increase maxusers.

Here are links for nproc and maxusers kernel parameters and the 11.0 Configurable Kernel parameter guide.


http://docs.hp.com/hpux/onlinedocs/os/KCparam.Nproc.html
http://docs.hp.com/hpux/onlinedocs/os/KCparam.MaxUsers.html
http://docs.hp.com/hpux/onlinedocs/os/KCparams.OverviewAll.html

[Kylie Spooner]

Thanks you so much Sri and Patrick for the fast reply...I think maybe you are right about the nproc as I did some reading on it earlier. My main problem right now though is how to regain control/access to my box. I was hoping not to reboot it. It has been up now for 230 days.
Ah the wonders of unix!

Melb