Doug McCallum wrote: > .... > What seemed potentially useful to us was if we could lock properties > that were operational properties like the actual path to use since there > are checks that must be done in order to have everything work correctly. > There are also value ranges on properties that could inadvertantly be set > to an inconsistent value if someone just changed them arbitrarily.
In reading this and thinking about this problem from a different angle, it almost seems like we have a new need for a method: one that can look at the current state of a service and say "no, you cannot enable me yet" or "yes I can be enabled" In the case of NFS, this would allow you to: - check properties as you've outlined above - check dfstab to see if anything is active therein - allow returning an abitrary string to svcadm for display I'm not eager to see the same replicated for disable because more often than not you will want what is required for the enable to still "be there" when you do the disable. Darren
