It has been a while since I have last dealt with an HP-UX server. Therefore, my memory is not too fresh and I want to consult the public wisdom for this.
I have a task at hand to come up with a measurable monitoring criteria for HP-UX servers, specific to this OS (version 11.X where X can be 0, 11, 11i, 23 or 31) and specific to HP hardware. The general monitoring rules are set by default, such as if CPU utilization is above 95% or disk utilization is above certain threshold, or someone is unsuccessfully trying to login to some unauthorized accounts. These and alike conditions are valid and trivial to monitor for any UNIX flavor. I am not interested in those. I am interested in things which might break in HP-UX and goes unnoticed unless someone specifically looks for them. One example :
if [ $(ioscan -fn | grep NO_HW | wc -l) -gt 0 ] then echo "One or more device(s) went missing" fi
This is a very simple one, I know. Simplicity is the reason why I can remember. Also, since
ioscan is not available on any other UNIX flavor, it is very HP-UX specific.
There must be other things that you can monitor on HP-UX only and I want to compile a list of such criteria. Such as any critical but non-obvious processes that needs to run all the time, comes to mind. Also I remember some magic strings you could feed to
cstm command, to inquire about your CPU, memory or other hardware.
If you have such esoteric commands and you're willing to share, it would be much appreciated.