[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: snmpconf Issue #18 resolution - Update capabilities table


I hope that we can agree that this is a potentially difficult problem.
However, in the case that you bring up, I believe that one has to run
a classifier program that probes the managed systems. Of course if you
have 100 systems, and 30% are brand X, 45% are brand Y, and the remaining
25% are brand Z, you don't need to do extensive probing of each of
the 100 systems for each new management app.

I get a little worried about deploying a new management app (or script)
on 100 systems at the same time on a live network that I have not tested
pretty extensively on a single system in a controlled environment.

A capabilities mechanism that allows me to skip testing,
but will fail sometimes (because of approximate completion
and low precision in its description of capabilities) 
doesn't seem like a prudent addition.
I've still got to test in my world.

At 03:59 PM 4/20/2001 -0400, Joel M. Halpern wrote:
>In your description Dave, you state that the management app should "probe the capabilities of the individual device.  Given taht what we are trying to do is provide a means my which a common set of configuration can be defined by the administrator, and applied across a number of devices, there is no way that the application can "know" much about what capabilities are required unless the language / libraries / objects tell it so.  What we are trying to define are the hooks that make it likely that the policy will be applicable / useful.  We recognize that you can not define fine enough grained capabilities in a general fashion.  Instead we are trying to provide coarse grained information that will help the management system determine whether a given configuration policy is likely to be useful to send to a given system.
>Joel M. Halpern

/david t. perkins