[olug] linux web server management ?

[William E. Kempf]

Phil Brutsche said:
> A long time ago, in a galaxy far, far way, someone said...
>
>> I'm not familiar with Debian or debconf in particular.  Does debconf
>> provide a full interface for managing configuration information?  In
>> other words, can I use it over and over to tweak an application's
>> configuration, with full access to all configurable options?
>
> No.
>
>> If so, that's the sort of solution I think is necessary, but it has to
>> be portable (i.e. I should be able to _easily_ use it on my Mandrake
>> box) and should be based on standard interfaces, rather than hacked
>> code for every package.  With a good set of standards you could
>> provide a generic utility for configuring packages that don't supply a
>> more robust utility.
>
> Good luck!
>
> This has been suggested over and over and over....

Then why no solution yet ;).  (This is a more than half serious question.)

> Until all developers reach a consensus as to what standard configuration
> system to use, you will constantly be learning different config files
> and writing custom modules for stuff like Webmin.

Uh huh.  But that's what standards are for.  The same could have been said
for the various other Linux standardization efforts, such as for directory
structure and usage, but developers *ARE* moving towards those standards
because it's a good idea.  If someone really did do the work to define a
_good_ standard here, you'd find developers moving towards this.

> Windows - the most common example - isn't as simple as people suggest.

Maybe you think they are suggesting that it's simpler than they really are?

> On the few Windows servers I am responsible for (or have been
> responsible for) at work I have different pieces of software to
> configure:
>
> Active Directory
> NT4 Domains
> IIS
> SMTP/POP3/IMAP servers
> database servers
> fax server software
>
> Each and every one of them requires a different tool - some of them
> aren't all that well documented compared to the Apache documentation web
> site or the Exim specification.  Being a GUI doesn't make the
> configuration tool any simpler :(

I don't agree.  A GUI is almost always simpler to use than a flat text
file.  This is usually due to a psychological factor... programmers view
config files as an implementation detail, and so don't put the same
emphasis on making its interface usable as they would a GUI.  But what
ever the reason, I've rarely run into a GUI config tool that I couldn't
make use of with out ever consulting the documentation.  I certainly can't
say the same thing for editing a config file.

> The only way to make everything configurable through the same interface
> is to throw out anything not written by Microsoft and abandon everything
> older than Windows 2000.

Uh huh.  Most of those things you listed above put their configuration
information in the registry, so I *CAN* use a single tool, regedit, to
configure them all.  Granted, this would be a horrible way to configure
things on Windows, as the registry isn't designed for this.  But enforcing
a certain structure on the general heirarchy, and including meta data
describing the final configuration options, and you get something that
*CAN* have a universal configuration tool.  XML and DTDs, for instance,
should be able to achieve this.  You'll still be much better off with a
custom config program (or plugin would be better, so we have a single
config tool), but at least there's a chance of being able to configure
something for which the developers have failed to give you a custom
solution.

>> Better yet, with standards established you can present APIs for
>> accessing configuration information, removing this burden entirely
>> from individual applications.  In other words, from a developer's
>> stand point the Win32 Registry API is a god send, even if the
>> monolothic binary registry file is obviously a mistake that Linux
>> wouldn't need to repeat to get the benefits of a standard
>> configuration mechanism.
>
> This has been suggested - the KDE and GNOME desktop environments even
> use it.  And it's a WONDERFUL idea.

I'm not familiar with any of these APIs, so I can't comment too much.  But
if it's being done, it's failing because there's no standard behind it (or
more specifically, they only address KDE or GNOME specific apps).

> But it's just that - an idea - until you can get people to actually USE
> the APIs.

I truly don't think this should be a problem if you define the APIs etc.
well enough to be universally (or nearly universally) usable.  I would
expect developers to rejoice in such an offering (and I *am* a developer).

William E. Kempf