>> I'm looking to prove that by using a zone I can increase the overall
>> stability of a server (if the app falls over it can't the the machine with
>> it).
> A kernel panic is a kernel panic is a kernel panic. :-)
> So if an app crashes the complete zone, chances are it will crash the
> entire box.  After all, it's only one Solaris kernel running.
>> Does anyone know of a program/script that would bring down a server if it
>> was running in the glabal zone, but wouldn't take the system down if it ran
>> in a non-global zone? 
> Out of idle curiosity, it would be interesting to find such a program
> or script.  But I am afraid it would not prove much.
> Sorry to sound so discouraging.  Zones are good for sandboxing apps,
> sharing hardware in a transparent way, providing uniform service
> containers for higher-level management (e.g. cluster), for localizing
> namespaces (uid/gid; /etc/services & ports, etc.) and many other things.
> Just not for kernel-crash-proofing the physical server.
> Regards -- Volker

Thanks Volker,

Out of curiosity then, how much I induce a kernel panic?  Now I've gotten this 
far into my investigation, I need to prove, or dis-prove that zones will all 
more stability.  If I can induce a kernel panic I can say that my initial 
thought was incorrect.

Many thanks,
This message posted from opensolaris.org
zones-discuss mailing list

Reply via email to