(new/old) idea of resolving

[Martin Kuzela <ku...@adela.sk>]

I know that resolving of IP address is time consuming and thats why
there is option to switch online resolving off and resolve addresses
later. But later
there is much less chance to resolve address at all.

What about write some background process with responsibility of writing
logs
and resolving IP addresses in them on background?

any problem?

Martin Kuzela

Re: (new/old) idea of resolving

[Rasmus Lerdorf <ra...@apache.org>]

> On Tue, Sep 07, 1999 at 10:32:52AM -0400, Rasmus Lerdorf wrote:
> > > I know that resolving of IP address is time consuming and thats why
> > > there is option to switch online resolving off and resolve addresses
> > > later. But later
> > > there is much less chance to resolve address at all.
> > 
> > Why?  DNS is not a transient thing.
> 
> On the contrary, it is very very transient. The PTR records can change
> on a regular basis. In fact with dynamic DNS, the PTR record could be
> changed (and on some networks are) when a dial-up user connects or
> diconnects.

Is that really worth worrying about?  Regardless, Apache supports piped
logs.  You can pipe the log entries to any program you want and that
program can resolve stuff.

-Rasmus

Re: (new/old) idea of resolving

[Peter Galbavy <Pe...@knowledge.com>]

On Tue, Sep 07, 1999 at 10:32:52AM -0400, Rasmus Lerdorf wrote:
> > I know that resolving of IP address is time consuming and thats why
> > there is option to switch online resolving off and resolve addresses
> > later. But later
> > there is much less chance to resolve address at all.
> 
> Why?  DNS is not a transient thing.

On the contrary, it is very very transient. The PTR records can change
on a regular basis. In fact with dynamic DNS, the PTR record could be
changed (and on some networks are) when a dial-up user connects or
diconnects.

Regards,
-- 
Peter Galbavy
Knowledge Matters Ltd
http://www.knowledge.com/

Re: (new/old) idea of resolving

[David Reid <ab...@dial.pipex.com>]

Where I sometimes work they have a problem with using dynamic IP allocation
and a desire to review what users have accessed from the internet via their
proxy.  Having the proxy lookup every IP slows the thing to a crawl so they
can't have it on, but without it they can't track who downloaded what!

Might be nice if Apache did offer the functionality, whether as a module or
as a support program using the piped logs?

Just my 2p worth.

david
----- Original Message -----
From: Dirk-Willem van Gulik <di...@webweaving.org>
To: <ne...@apache.org>
Sent: 07 September 1999 16:08
Subject: Re: (new/old) idea of resolving


>
> On Tue, 7 Sep 1999, Rasmus Lerdorf wrote:
>
> > Why?  DNS is not a transient thing.
>
> Though some idiots^H^H^H^Hclever people are starting to do things like
> using quick DNS updates to for example assign very short lived reverse
> mappings to constantly changing dhcp or ppp addresses.
>
> Dynamic DNS Update is a mixed blessing.
>
> I.e. while a customer is on line, or while a machine has been dhcp-ed in,
> its reverse address it is nicely called 'customername.we.trust.you.com' or
> somthing like that. After log-off the mapping is removed.
>
> I've had two US customers who wanted me to implement this for them, tied
> in with security and xs control based on the short lived fqhn.
>
> And looking at the next windows release, that is what some security seems
> to start rely on. Going be nice to see the battle over which IT staff gets
> to run the authoritative DNS servers; as for some things you might really
> want to keep it close/on your NT domain controller.
>
> So there is some argument for resolving relatively quickly; though of
> course some DNS implementations their caches are hopelessly broken when
> the TTL < 15 minutes. So it cannot be that quick. Yet.
>
> Dw.
>

Re: (new/old) idea of resolving

[Dirk-Willem van Gulik <di...@webweaving.org>]

On Tue, 7 Sep 1999, Rasmus Lerdorf wrote:

> And personally I hope it doesn't.

+1.

Dw

Re: (new/old) idea of resolving

[Rasmus Lerdorf <ra...@apache.org>]

> So there is some argument for resolving relatively quickly; though of
> course some DNS implementations their caches are hopelessly broken when
> the TTL < 15 minutes. So it cannot be that quick. Yet.

And Solaris ships with nscd which by default has a one hour TTL on its
cache.  I think it is going to be a while before short-expiry dns like
this becomes useful.  And personally I hope it doesn't.

-Rasmus

Re: (new/old) idea of resolving

[Dirk-Willem van Gulik <di...@webweaving.org>]

On Tue, 7 Sep 1999, Rasmus Lerdorf wrote:

> Why?  DNS is not a transient thing.

Though some idiots^H^H^H^Hclever people are starting to do things like
using quick DNS updates to for example assign very short lived reverse
mappings to constantly changing dhcp or ppp addresses.

Dynamic DNS Update is a mixed blessing. 

I.e. while a customer is on line, or while a machine has been dhcp-ed in,
its reverse address it is nicely called 'customername.we.trust.you.com' or
somthing like that. After log-off the mapping is removed.

I've had two US customers who wanted me to implement this for them, tied
in with security and xs control based on the short lived fqhn. 

And looking at the next windows release, that is what some security seems
to start rely on. Going be nice to see the battle over which IT staff gets
to run the authoritative DNS servers; as for some things you might really
want to keep it close/on your NT domain controller.

So there is some argument for resolving relatively quickly; though of
course some DNS implementations their caches are hopelessly broken when
the TTL < 15 minutes. So it cannot be that quick. Yet.

Dw.

Re: (new/old) idea of resolving

[Rasmus Lerdorf <ra...@apache.org>]

> I know that resolving of IP address is time consuming and thats why
> there is option to switch online resolving off and resolve addresses
> later. But later
> there is much less chance to resolve address at all.

Why?  DNS is not a transient thing.

-Rasmus