>Daring to challenge the will of the almighty Leviam00se, matthew c. mead
>: Hi - I'm currently attempting to get amd to read an NIS
>: (really Solaris 2.x NIS+ in NIS compat mode - yeah yeah, I know)
>No you don't. If you really knew, you'd understand exactly why this
>is happening. :)
If you'd read your mail, clearly you'd know I do.
Quote:>The Sun rpc.nisd in YP compat mode does not support the YPPROC_ORDER
>procedure from NIS v2. (Don't ask me why: Sun just did it that way.)
>This can hose amd because it uses yp_order() to check the last update
>time of its maps to see whether or not it needs to reload them.
Yes, yes, didn't you explain this once before? Or do you
like to see your posts?
Quote:>Unfortunately, there's no way to work around this in FreeBSD's NIS client
>code: without the YPPROC_ORDER procedure on the server end, yp_order()
>will always return an error. The only hack (without having NIS+ available)
>is to modify amd to periodically reload its maps regardless of what
>yp_order() says.
I know for certain there are workarounds - by
incorporating nisv3 code into the libs (or at least amd's lib) so
that you can use an "NIS+" call to determine map age. One would
think that interoperability with other operating systems (in this
case, Solaris) would be one of FreeBSD's goals. Without said
interoperability, what good is the OS?
Quote:>You might also consider installing the Sun NS Kit on the Solaris box
>so that you can use a real ypserv (or set up a FreeBSD ypserv; you
>could set up a separate domain just to hold your amd maps and host it
>from a FreeBSD box). I also think Solaris 2.6 comes with ypserv bundled.
Well, it's pretty clear I can do that. It seems like
such a waste, though, to have a dedicated machine to distribute
amd maps. I don't know for sure whether or not an NIS server and
and NIS+ server in NIS compatibility mode would coexist on a
Solaris machine, but I'd wager not. Again, a second dedicated
machine. Clearly, hacking amd to just reload the *y map
every 10 minutes is preferable to this sort of resource waste.
-matt
--
Matthew C. Mead
http://www.veryComputer.com/~mmead/