[Postgres-xc-general] postgres-xc 0.9.7 : failed to connect to data node/can not connect to datanode 0

[Phil S. <phi...@ne...>]

Hello all ...

I am in the process of setting up and running postgres-XC 0.9.7 on 
Ubuntu 10.04 LTS 64bit in a development/test sandbox.

I used the install manual, 
http://iweb.dl.sourceforge.net/project/postgres-xc/Version_0.9.7/PG-XC_InstallManual_v0_9_7.pdf 
<http://iweb.dl.sourceforge.net/project/postgres-xc/Version_0.9.7/PG-XC_InstallManual_v0_9_7.pdf>, 
to set up one GTM VM host and two coordinator/datanode VM hosts as this 
is the recommended environment per the manual.

After I initialized the node information in each coordinator using the 
Create Node SQL, I am getting the following error trying to create a 
database on the coord1 node:

     LOG:  failed to connect to data node
     WARNING:  can not connect to datanode 0
     LOG:  failed to acquire connections
     STATEMENT:  create database TEST;
     ERROR:  Failed to get pooled connections
     STATEMENT:  create database TEST;

Was wondering if someone could suggest what is causing my connection 
issues.  I have searched through all the forums on sourceforge as well 
as the net and haven't been able to find anything useful yet.

Thanks in advance,
Phil.




------------------------------


Here are the details:


GTM host install:
----------------
as root:
- installed all necessary Ubuntu packages required for postgres-xc install
- created the postgres-xc user
- created a postgres-xc build directory
- downloaded pgxc_v0.9.7.tar.gz from sourgeforge.net 
<http://sourgeforge.net> and untarred it in build dir above
- config'ed postgres-xc build pointing to custom db home directory with 
NLS enabled and perl
- ran 'make' and 'make install' without issues
- created custom gtm data and log directories

as postgres-xc:
- copied gtm.conf.sample to gtm data dir
- edited the file as per the install manual:

     nodename = 'one'
     listen_addresses = '*'
     port = 6666
     log_file = '/home/postgres/gtm/log/gtm.log'

- started the GTM without issue using the following cmd:

     gtm_ctl start -S gtm -D /home/postgres/gtm/data -l 
/home/postgres/gtm/log/gtm.log


Coord/Datanode host install:
---------------------------

as root:
- installed all necessary Ubuntu packages required for postgres-xc install
- created the postgres-xc user
- created a postgres-xc build directory
- downloaded pgxc_v0.9.7.tar.gz from sourgeforge.net 
<http://sourgeforge.net> and untarred it in build dir above
- config'ed postgres-xc build pointing to custom db home directory with 
NLS enabled and perl
- ran 'make' and 'make install' without issues
- created custom coordinator and datanode data and log directories


as postgres-xc:
- ran initdb for cooordinator on each of the two db hosts
   - data directory was customized (ie not default)
   - nodename was coord1/coord2 respective of db_host_1/db_host_2

- ran initdb for datanode on each of the two db hosts
   - data directory was customized (ie not default)
   - nodename was data1/data2 respective of db_host_1/db_host_2

- edited coordinators postgres.conf and pg_hba.conf respective of the db 
host
   - postgresql.conf
     - pooler_port=6667            (just commented it out)
     - gmt_host=<IP address of GTM host>    (just commented it out)
     - gmt_post=6666                (just commented it out)
     - pgxc_node_name=coord1         (didn't have to change this)  ---> 
it was set to coord2 on the 2nd db host
     - port=5432                (just commented it out)
   - pg_hba.conf
     - added the following line to the end of the file:
         host    all    all <IP_subnet>.0/32    trust    ---> all hosts 
are on the same IP_subnet


- edited datanodes postgres.conf and pg_hba.conf respective of the db host
   - postgresql.conf
     - gmt_host=<IP address of GTM host>    (just commented it out)
     - gmt_post=6666                (just commented it out)
     - pgxc_node_name=data1             (didn't have to change this)  
---> it was set to data2 on the 2nd db host
     - port=15432                (edited it to change to 15432)
   - pg_hba.conf
     - added the following line to the end of the file:
         host    all    all <IP_subnet>.0/32    trust    ---> all hosts 
are on the same IP_subnet

- started the coordinator on each node without issue using the folowing 
command:

     pg_ctl start -D /home/postgres/coordinator/data -l 
/home/postgresql/coordinator/log/coordinator1.log -Z coordinator  ---> 
log is coordinator2 on 2nd db host


- started the datanode on each node without issue using the folowing 
command:

     pg_ctl start -D /home/postgres/datanode/data -l 
/home/postgresql/coordinator/log/datanode1.log -o "-p 15432" -Z 
datanode  ---> log is datanode2 on 2nd db host



Node inititialization on each coordinator:
-----------------------------------------

logged onto each coordinator using psql and ran the following cmds:

     psql -U postgres -d postgres

     CREATE NODE coord1 WITH (TYPE = 'coordinator', PORT = 5432, HOST = 
'<IP_of_coord1>');
     CREATE NODE coord2 WITH (TYPE = 'coordinator', PORT = 5432, HOST = 
'<IP_of_coord2>');

     CREATE NODE data1 WITH (TYPE = 'datanode', PORT = 15432, HOST = 
'<IP_of_data1>');
     CREATE NODE data2 WITH (TYPE = 'datanode', PORT = 15432, HOST = 
'<IP_of_data2>');

     SELECT pgxc_pool_reload();