[Pgpool-general] Authentication method used for sr_check_password, health_check_password and recovery_password

[Lazaro Rubén García Martinez]

I am agree with you, but if it is not a bug, what is the purpose for having sr_sheck_password property in pgpool.conf file?.

I think this property can confuse pgpool's users, for this reason I propose -1.

If you understand that this feature should be present in Pgpool 3.2, I will agree with you too.

Regards.
________________________________________
De: Guillaume Lelarge [guillaume at lelarge.info]
Enviado el: domingo, 20 de noviembre de 2011 17:58
Para: Lazaro Rubén García Martinez
CC: Tatsuo Ishii; pgpool-general at pgfoundry.org
Asunto: RE: [Pgpool-general] Authentication method used for sr_check_password, health_check_password and recovery_password

On Sun, 2011-11-20 at 17:24 -0430, Lazaro Rubén García Martinez wrote:
> I think this feature is very important, because having  trust acces in pg_hba.conf is not a good idea.

I understand that and I agree with you. The problem is not on the
feature itself, but on which release it should be delivered. If the
feature is really urgent to get out there, then we should release 3.2
quickly. We shouldn't put it in 3.1.whatever because 3.1.whatever could
get out before 3.2.

Minor releases shouldn't change behaviour apart from bugfixes. That's an
important part of the trust you can have in a software. If we start to
add features on bugfix releases, many people will stop doing minor
updates on pgpool, afraid of bugs which might be included with new
features. I know I'll do if this will happen, and I won't encourage my
customers to upgrade their pgpool.

So, definite +1 to add this feature to pgpool, +1 to add it to 3.2, -1
to add it as a bugfix in 3.1.1. It definitely is not a bugfix.


--
Guillaume
  http://blog.guillaume.lelarge.info
  http://www.dalibo.com