Re: Add statistics_collector_listen_addresses to fix hard-coding of "localhost" - Mailing list pgsql-bugs

From Craig Ringer
Subject Re: Add statistics_collector_listen_addresses to fix hard-coding of "localhost"
Date
Msg-id 4EBB3443.2020709@ringerc.id.au
Whole thread Raw
In response to Re: Add statistics_collector_listen_addresses to fix hard-coding of "localhost"  (Robert Young <yayooo@gmail.com>)
Responses Re: Add statistics_collector_listen_addresses to fix hard-coding of "localhost"  (Robert Young <yayooo@gmail.com>)
List pgsql-bugs
On 10/11/11 02:24, Robert Young wrote:
> Sure! It give the right information to workaround the problem.
> But as I said:
> On Fri, Oct 28, 2011 at 07:11, Robert Young <yayooo@gmail.com> wrote:
>> 2.hard-coding is NOT configurable, that is the problem must be aided
>> from OUTSIDE of the system to workaround.
>>
> And
> On Tue, Nov 1, 2011 at 08:04, Robert Young <yayooo@gmail.com> wrote:
>> Database should be functional without underlying hostname or DNS facility.

Why?

Should it also be functional without sockets?

How about without a file system?

Without an OS?


PostgreSQL is an application running on an OS, and expects to be able to
use the basic services of that OS. Those services include local DNS
resolution of local host names. Every OS since Windows 3.1.1, Mac OS 7,
and early BSD UNIXes has provided local DNS resolution and loopback
sockets. Why should PostgreSQL accomodate the lack of them? If it
should, where does it stop accomodating broken and cut-down OSes?

If you want an embedded database for an extremely cut down OS,
PostgreSQL isn't particularly suitable for a lot of reasons. For any
other use case there's no reason to expect local DNS resolution to be
broken.

--
Craig Ringer

pgsql-bugs by date:

Previous
From: Robert Young
Date:
Subject: Re: Add statistics_collector_listen_addresses to fix hard-coding of "localhost"
Next
From: Robert Young
Date:
Subject: Re: Add statistics_collector_listen_addresses to fix hard-coding of "localhost"