[bind10-dev] #1292 and data source configuration

[Jelte Jansen]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 11/24/2011 12:44 AM, JINMEI Tatuya / 神明達哉 wrote:
> At Wed, 23 Nov 2011 13:07:18 +0100,
> Jelte Jansen <jelte at isc.org> wrote:
> 
>>> This leads to a more generic topic of how we configure data sources
>>> (or data source (client) modules) system-wide.  I guess we'll
>>> eventually have some global configuration (i.e., config not specific
>>> to a particular process) like TSIG keys, to provide minimum
>>> information to create data source clients for a particular type of
>>> data source.  That would include a path to the corresponding loadable
>>> module:
> 
>> side note; We are getting so many 'global' options, should we start
>> grouping them?
> 
> Do we have many?  Do you specifically mean we now have "global" config
> for logging and and TSIG and are now possibly introducing for data
> source?
> 

ok, this would make three, and i think i've heard talk about one or two
others.

> In any case, I've not thought about that and right now don't have a
> strong opinion.  Maybe we can have a single top level category (at the
> same level as "Auth", "Resolver", etc) such as "System" where
> TSIG/logging/data source/etc configs belong.  Or maybe not.
> 

I didn't have a specific proposal yet either, just thinking aloud about it.

> 
> The example with "module_path" is just for example.  I'm not
> necessarily pushing it or don't even necessarily thought that was the
> best idea.  I actually don't have a strong preference, but maybe we
> should follow common convention (if any).  From a quick look, BIND 9's
> dlz just uses an absolute file name.  apache seems to have a hybrid
> style:
> LoadModule ssl_module libexec/apache22/mod_ssl.so
> The complete absolute path is constructed by prepending the value of
> ServerRoot to the specified relative path.  (I guess an absolute path
> can also be specified for LoadModule).
> 

That sounds similar to what I had in mind. One thing I was wondering is
whether we should support having multiple directories (one system, one
user-defined, maybe one source-if-run-from-sourcetree), where we
strictly define the search order (which is, after all, one of the
original problems here).

>> I agree that our call to dlopen() should use a full path (as obviously
>> different implementations use different algorithms for searching), but
>> depending on the resolution of the ticket above, I'm not sure whether
>> the middle-term way would need additional (and, more importantly,
>> temporary) values used in configuration.
> 
> I don't have a strong opinion on the specific solution in the context
> of #1292.  As long as it's not too ad hoc and not too generic (and
> possibly complicated) for the immediate need, I'm okay with any
> solution.
> 

We'll continue discussion on that in the ticket, and use this thread to
discuss the longer-term approach.

(Other people: feel free to chime in there too! and/or pull it back to
the list if you want)

Jelte
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk7OGugACgkQ4nZCKsdOncXDfwCgudcXlUz0PazOsylueKowwgPy
8roAoKXRsTx6DA3ojJkQ1+CtDqSV7Enb
=aub6
-----END PGP SIGNATURE-----