Solaris

I've often added entries to the /etc/project file, and it's always worked fine. However, I've run into an application that's been having resource problems, and I've been trying to research causes.

My question is how does rcapd relate to /etc/project? I see references to how it manages resource and uses /etc/project in various articles I've read, in the man pages, etc.... but I haven't found if rcapd is actually required for entries in /etc/project to actually work properly. In my environment svc:/system/rcap:default is normally disabled. Our DBAs certainly have never complained after entries have been added to /etc/project, and rcapd is still disabled.

thanks guys!

rcapd enforces limits on resources. Period. What problems? ...and what does problem mean in this case?

I suspect oracle....

I guess my confusion is where are these limits on resources configured when rcapd is enable? I had originally thought /etc/project and /etc/system were the only places where you managed resources for the kernel.

Thanks

You should check if you application is actually running under correct project:

Code:

ps -o project -p PID

Is rcapd actually correlated with /etc/project?

Thanks for that ps tip.

It is the other way around - project works under the "auspices" of rcapd.

rcapd behaves in response to part of zonecfg information - example partial "zonecfg info" output from a zone with capped cpus:

Code:

capped-cpu:
        [ncpus: 2.00]
rctl:
        name: zone.cpu-cap
        value: (priv=privileged,limit=200,action=deny)

I thought zone resource management was handled strictly through Fair Share Scheduler (FSS).

I guess I'm still confused on how /etc/project does or doesn't correlate with rcapd. I can't seem to find any good documentation on rcapd.

Thanks for the help!