pg_ctl wrong status info on Windows - Mailing list pgsql-bugs

From Lucian Curelaru
Subject pg_ctl wrong status info on Windows
Date
Msg-id CAGLKO7zcxVEoAdfsh7GoFrXf0LRtpbCpwBTqTeqGaTwQ5A4G2Q@mail.gmail.com
Whole thread Raw
List pgsql-bugs
Hi

I have a Postgresql 9.1.2 installed on a Windows 7 machine and noticed
the following thing:

- I start the postgresql by windows service ( on default port )
- when I run pg_ctl -D on the same data directory the response is "no
server running" even the server is in fact running already
- if I try to start using pg_ctl start ...   I receive the error that
the port is occupied ( correct )
- if I try to start it on a diffrent port.. the start is succesful and
I'm able to run two servers on same data via two diffrent ports. I've
connected to them and modified diffrent things in the two databases :)
and the databases behived like diffrent ones until the next restart of
both servers when inserted data has mixed and messed up ( partially
was lost / partially mixed up)
Anyway.. back to the original issue it seems not ok from my point of
view that pg_ctl seems not to be aware of the postgres instance
started by the windows service. I've tried also to first start server
from pg_ctl and then from service.. and the service tried to start and
then aborted leaving the current server to run alone (correct)
I've noticed also that the windows service is creating the pid file
but the pg_ctl seems to ignore it on status commands and overwrite it
on start commands.
On the other hand I've performed several tests (starting , stoping
,status servers on same datadir) only with pg_ctl and all seems to
work correctly.
Am I missing something or just find a bug?


Lucian

pgsql-bugs by date:

Previous
From: Hitoshi Harada
Date:
Subject: Re: BUG #6572: The example of SPI_execute is bogus
Next
From: Alex Matzinger
Date:
Subject: Re: BUG #6571: Postgres Kills 'Select 1' query randomly on hot standby databases