[Pgpool-general] pgpool transport endpoint connection error

[Tatsuo Ishii]

Hum. I suspect that the server both PostgreSQL and pgpool running on
runs out some resources related to networking(I'm not an networking
expert, so this is just a guess). I guess that's the reason why you
don't see errors when pgpool is not used.
--
Tatsuo Ishii
SRA OSS, Inc. Japan

> It's a PostgreSQL error, but only occurs when pgpool is used for the
> connection. If I connect to the server directly, without pgpool, it doesn't
> not give that error under any circumstance/load. Ideally, I'd like to
> continue using pgpool, so if some light could be shed as to how to resolve
> it, that would be excellent.
> 
> TIA
> 
> On 3/12/07, Tatsuo Ishii <ishii at sraoss.co.jp> wrote:
> >
> > Can I make things clear?
> >
> > Did you find the error on the PostgreSQL server or pgpool log?
> > --
> > Tatsuo Ishii
> > SRA OSS, Inc. Japan
> >
> >
> > > Thanks for the prompt response, however:
> > >
> > > $ grep "max_connections" /db/postgresql.conf
> > > max_connections = 600                   # (change requires restart)
> > > # Note: increasing max_connections costs ~400 bytes of shared memory per
> > > shared_buffers = 128MB                  # min 128kB or
> > max_connections*16kB
> > > # max_locks_per_transaction * (max_connections +
> > max_prepared_transactions)
> > >
> > > On 3/12/07, Tatsuo Ishii <ishii at sraoss.co.jp> wrote:
> > > >
> > > > > Hi,
> > > > >
> > > > > After searching the forum archives, I came across another post which
> > had
> > > > the
> > > > > same error message I am having:
> > > > >
> > > > > http://ml.postgresql.jp/pipermail/pgsql-jp/2004-February/015725.html
> > > >
> > > > Problem described above is not relevant to yours since they didn't use
> > > > pgpool.
> > > >
> > > > Besides that I'm not sure what your problem is coming from, though I
> > > > suspect your PostgreSQL's max_connections is lower than
> > > > num_init_children * (# of user/database pairs of your pgpool clients).
> > > > --
> > > > Tatsuo Ishii
> > > > SRA OSS, Inc. Japan
> > > >
> > > > > My error message reads as:
> > > > >
> > > > > SQLSTATE[08006] [7] could not send startup packet: Transport
> > endpoint is
> > > > not
> > > > > connected
> > > > >
> > > > > # pgpool -h
> > > > > pgpool version 3.2(kawakawa),
> > > > >   a generic connection pool/replication/load balance server for
> > > > PostgreSQL
> > > > >
> > > > > My config:
> > > > >
> > > > > listen_addresses = 'localhost'
> > > > > port = 5434
> > > > > socket_dir = '/tmp'
> > > > > backend_host_name = ''
> > > > > backend_port = 5432
> > > > > backend_socket_dir = '/tmp'
> > > > > secondary_backend_host_name = ''
> > > > > secondary_backend_port = 0
> > > > > num_init_children = 128
> > > > > max_pool = 16
> > > > > child_life_time = 0
> > > > > connection_life_time = 15
> > > > > child_max_connections = 0
> > > > > logdir = '/var/run'
> > > > > replication_mode = false
> > > > > connection_cache = true
> > > > > health_check_timeout = 20
> > > > > health_check_period = 0
> > > > > health_check_user = 'nobody'
> > > > > insert_lock = false
> > > > > ignore_leading_white_space = false
> > > > >
> > > > >
> > > > > I have about 170 open connections when it starts giving me the error
> > > > message
> > > > > on connect. Any ideas?
> > > > >
> > > > > TIA