Thread: strange connection problem.
I have a development site ( FreeBSD 10.1 and Postgres 9.3 ) where I can connect to the database via psql or pgadminIII both locally and remotely.
However, all of the local apps ( 3 different ones ) that would connect to it are failing to connect.
I see no error events in the messages/postgres log files related to bad attempts.
I suspect something with name resolution,
The hosts, hosts.conf and resolv.conf files are set to resolv first via the hosts file
And the hostname and localhost are properly defined there.
The site installation did have some initial DNS setup issues. However this command works fine locally.
“psql –U <role> -h <hostname> database.
i.e. name resolution works locally fine for psql.
The problem does not go away on shutting down and starting up the system ( Virtual Machine )
Is there anything on the postgres side that might explain this situation ?
Will likely re-install and see if the problem goes away. ( build of the day problem ? )
It is an interesting puzzle at the moment.
Thanks
Dave
On 2015-10-23 08:29, Day, David wrote: > I have a development site ( FreeBSD 10.1 and Postgres 9.3 ) where I can > connect to the database via psql or pgadminIII both locally and remotely. > > However, all of the local apps ( 3 different ones ) that would connect to it > are failing to connect. Sounds like your apps are using a different port than the default install. Use "psql -p xxxx" with the port from the URI configured in your apps. -- http://yves.zioup.com gpg: 4096R/32B0F416
I have a development site ( FreeBSD 10.1 and Postgres 9.3 ) where I can connect to the database via psql or pgadminIII both locally and remotely.
However, all of the local apps ( 3 different ones ) that would connect to it are failing to connect.
On 10/23/2015 07:29 AM, Day, David wrote: > I have a development site ( FreeBSD 10.1 and Postgres 9.3 ) where I can > connect to the database via psql or pgadminIII both locally and remotely. > > However, all of the local apps ( 3 different ones ) that would connect > to it are failing to connect. > > I see no error events in the messages/postgres log files related to bad > attempts. > > I suspect something with name resolution, > > The hosts, hosts.conf and resolv.conf files are set to resolv first via > the hosts file > > And the hostname and localhost are properly defined there. > > The site installation did have some initial DNS setup issues. However > this command works fine locally. > > “psql –U <role> -h <hostname> database. > > i.e. name resolution works locally fine for psql. > > The problem does not go away on shutting down and starting up the system > ( Virtual Machine ) > > Is there anything on the postgres side that might explain this situation ? What are the connection parameters for the local apps versus the ones you used above for psql? In particular the user? What is in your pg_hba.conf file? When you say locally are you talking about everything happening in the VM or between the host machine and the VM or some combination thereof? What are your logging parameters in postgresql.conf? > > Will likely re-install and see if the problem goes away. ( build of the > day problem ? ) > > It is an interesting puzzle at the moment. > > Thanks > > Dave > -- Adrian Klaver adrian.klaver@aklaver.com
One app is connecting through libpqxx, the other 2, glassfish and ejabber chat agent, I am not certain.
It is only on this one machine where I am having a connection issue with any of these apps.
I can connect via the identical roles that these apps would be using via the psql command line on this system.
I doubt that all three would have a connection issue on the same day, without some kind of global matter.
My pg_hba.conf file on this system is the following.
# "local" is for Unix domain socket connections only
local all all trust
# IPv4 local connections:
host all all 127.0.0.1/32 trust
# IPv6 local connections:
host all all ::1/128 trust
# Allow replication connections from localhost, by a user with the
# replication privilege.
host all all 192.168.32.0/19 trust
I don’t believe this to be a postgres problem, but you don’t know what you don’t know.
From: David G. Johnston [mailto:david.g.johnston@gmail.com]
Sent: Friday, October 23, 2015 10:36 AM
To: Day, David
Cc: pgsql-general@postgresql.org
Subject: Re: [GENERAL] strange connection problem.
On Fri, Oct 23, 2015 at 10:29 AM, Day, David <dday@redcom.com> wrote:
I have a development site ( FreeBSD 10.1 and Postgres 9.3 ) where I can connect to the database via psql or pgadminIII both locally and remotely.
However, all of the local apps ( 3 different ones ) that would connect to it are failing to connect.
So how do the methods the local apps use to connect to PostgreSQL differ compared to running psql directly on the same machine?
David J.
"Day, David" <dday@redcom.com> writes:
> I have a development site ( FreeBSD 10.1 and Postgres 9.3 ) where I can connect to the database via psql or
pgadminIIIboth locally and remotely.
> However, all of the local apps ( 3 different ones ) that would connect to it are failing to connect.
Are those apps trying to use TCP connections, or Unix-socket connections?
If the latter, it might be a discrepancy in where they expect the socket
file to be versus where the postmaster thinks it should be.
regards, tom lane
On 10/23/2015 08:23 AM, Day, David wrote: > One app is connecting through libpqxx, the other 2, glassfish and > ejabber chat agent, I am not certain. > > It is only on this one machine where I am having a connection issue > with any of these apps. > > I can connect via the identical roles that these apps would be using via > the psql command line on this system. > > I doubt that all three would have a connection issue on the same day, > without some kind of global matter. > > My pg_hba.conf file on this system is the following. > > # "local" is for Unix domain socket connections only > > local all all trust > > # IPv4 local connections: > > host all all 127.0.0.1/32 trust > > # IPv6 local connections: > > host all all ::1/128 trust > > # Allow replication connections from localhost, by a user with the > > # replication privilege. > > host all all 192.168.32.0/19 trust > > I don’t believe this to be a postgres problem, but you don’t know what > you don’t know. > Can you launch one or more of the apps from the command line and see if any errors appear there, that are not bubbling up to the app? Also, and to answer Toms questions, what are the exact connection parameters being used in the app(s)? -- Adrian Klaver adrian.klaver@aklaver.com
I figured this out and it is not a Postgres issue.
There was a area of initialization overlooked by the installer on a fresh install
of a development system..
I had to add a little extra debug to see that one key job not
only was not getting attached to the database, it was not even trying.
After seeing that I able to run down the installation flaws.
Thanks very much to all for your suggestions on this matter.
Regards
Dave
-----Original Message-----
From: Tom Lane [mailto:tgl@sss.pgh.pa.us]
Sent: Friday, October 23, 2015 12:49 PM
To: Day, David
Cc: pgsql-general@postgresql.org
Subject: Re: [GENERAL] strange connection problem.
"Day, David" <dday@redcom.com> writes:
> I have a development site ( FreeBSD 10.1 and Postgres 9.3 ) where I can connect to the database via psql or
pgadminIIIboth locally and remotely.
> However, all of the local apps ( 3 different ones ) that would connect to it are failing to connect.
Are those apps trying to use TCP connections, or Unix-socket connections?
If the latter, it might be a discrepancy in where they expect the socket file to be versus where the postmaster thinks
itshould be.
regards, tom lane