Question about "ident_file" in postgres.conf

On Tue, Nov 6, 2012 at 9:08 AM, Tianyin Xu <tixu@cs.ucsd.edu> wrote:

Hi, all,

I have a question regarding the "ident_file" configuration entry.

Why the server refused to start without specifying the "ident_file", but
it never cares whether the given "ident_file" is valid or not? In other
word, a misconfigured non-existent path for "ident_file" can also start the
server with a background message.

This does not make too much sense to me. I think the system behavior
should be opposite. If the user didn't specify the "ident_file", the pg
server should ignore. But if the user specifies one and that one is not
valid, the user should stop because it clearly means the path is
misconfigured.

Is there any concern here? Otherwise I would suggest and provide a patch
to check the validity of the "ident_file" like what pg is doing for
"data_directory" and "hba_file".

This is definitely a known problem. It's actually on the TODO list already,
just not phrased in a way that makes it likely to be found.

It should be treated the same way as the hba_file. So sure, please do
provide a patch for that if you can.

--
Magnus Hagander
Me: http://www.hagander.net/
Work: http://www.redpill-linpro.com/

On Tue, Nov 6, 2012 at 3:16 AM, Magnus Hagander <magnus@hagander.net> wrote:

On Tue, Nov 6, 2012 at 9:08 AM, Tianyin Xu <tixu@cs.ucsd.edu> wrote:

Hi, all,

I have a question regarding the "ident_file" configuration entry.

Why the server refused to start without specifying the "ident_file", but
it never cares whether the given "ident_file" is valid or not? In other
word, a misconfigured non-existent path for "ident_file" can also start the
server with a background message.

This does not make too much sense to me. I think the system behavior
should be opposite. If the user didn't specify the "ident_file", the pg
server should ignore. But if the user specifies one and that one is not
valid, the user should stop because it clearly means the path is
misconfigured.

Is there any concern here? Otherwise I would suggest and provide a patch
to check the validity of the "ident_file" like what pg is doing for
"data_directory" and "hba_file".

This is definitely a known problem. It's actually on the TODO list already,
just not phrased in a way that makes it likely to be found.

It should be treated the same way as the hba_file. So sure, please do
provide a patch for that if you can.

A patch for this is already committed for 9.3.

https://commitfest.postgresql.org/action/patch_view?id=889

Cheers,

Jeff

On Tue, Nov 6, 2012 at 5:25 PM, Jeff Janes <jeff.janes@gmail.com> wrote:

On Tue, Nov 6, 2012 at 3:16 AM, Magnus Hagander <magnus@hagander.net>
wrote:

On Tue, Nov 6, 2012 at 9:08 AM, Tianyin Xu <tixu@cs.ucsd.edu> wrote:

Hi, all,

I have a question regarding the "ident_file" configuration entry.

Why the server refused to start without specifying the "ident_file", but
it never cares whether the given "ident_file" is valid or not? In other
word, a misconfigured non-existent path for "ident_file" can also start

the

server with a background message.

This does not make too much sense to me. I think the system behavior
should be opposite. If the user didn't specify the "ident_file", the pg
server should ignore. But if the user specifies one and that one is not
valid, the user should stop because it clearly means the path is
misconfigured.

Is there any concern here? Otherwise I would suggest and provide a patch
to check the validity of the "ident_file" like what pg is doing for
"data_directory" and "hba_file".

This is definitely a known problem. It's actually on the TODO list

already,

just not phrased in a way that makes it likely to be found.

It should be treated the same way as the hba_file. So sure, please do
provide a patch for that if you can.

A patch for this is already committed for 9.3.

https://commitfest.postgresql.org/action/patch_view?id=889

D'oh. I knew that. Can't believe I forgot. Apologies!

--
Magnus Hagander
Me: http://www.hagander.net/
Work: http://www.redpill-linpro.com/

I see. Thanks a lot for the explanation!

Tianyin

On Tue, Nov 6, 2012 at 8:32 AM, Magnus Hagander <magnus@hagander.net> wrote:

On Tue, Nov 6, 2012 at 5:25 PM, Jeff Janes <jeff.janes@gmail.com> wrote:

On Tue, Nov 6, 2012 at 3:16 AM, Magnus Hagander <magnus@hagander.net>
wrote:

On Tue, Nov 6, 2012 at 9:08 AM, Tianyin Xu <tixu@cs.ucsd.edu> wrote:

Hi, all,

I have a question regarding the "ident_file" configuration entry.

Why the server refused to start without specifying the "ident_file",

but

it never cares whether the given "ident_file" is valid or not? In other
word, a misconfigured non-existent path for "ident_file" can also

start the

server with a background message.

This does not make too much sense to me. I think the system behavior
should be opposite. If the user didn't specify the "ident_file", the pg
server should ignore. But if the user specifies one and that one is not
valid, the user should stop because it clearly means the path is
misconfigured.

Is there any concern here? Otherwise I would suggest and provide a

patch

to check the validity of the "ident_file" like what pg is doing for
"data_directory" and "hba_file".

This is definitely a known problem. It's actually on the TODO list

already,

just not phrased in a way that makes it likely to be found.

It should be treated the same way as the hba_file. So sure, please do
provide a patch for that if you can.

A patch for this is already committed for 9.3.

https://commitfest.postgresql.org/action/patch_view?id=889

D'oh. I knew that. Can't believe I forgot. Apologies!

--
Magnus Hagander
Me: http://www.hagander.net/
Work: http://www.redpill-linpro.com/

--
Tianyin XU,
http://cseweb.ucsd.edu/~tixu/