Re: [Postgres-xc-general] coordinator deaths

[Michael P. <mic...@gm...>]

Hi,

You were not registered on the general mailing list, so I had to check the
content of your message manually.
This explains the delay of your message.

By the way, regarding your message, please see comments below.

On Wed, Sep 19, 2012 at 4:51 AM, Roger Mayes <rog...@gm...>wrote:

> I'm using CentOS 6.3, I've used the "short version" installation to set up
> a gtm, a single coordinator, and a single data node on my local host.  Here
> are the contents of my pgxc_node table:
>
> -bash-4.1$ psql -U postgres
> psql (PGXC 1.0.1, based on PG 9.1.5)
> Type "help" for help.
> postgres=# select * from pgxc_node;
>  node_name | node_type | node_port | node_host | nodeis_primary |
> nodeis_preferred |  node_id
>
> -----------+-----------+-----------+-----------+----------------+------------------+------------
>  coord1    | C         |      5432 | localhost | f              | f
>          | 1885696643
>  datanode1 | D         |     15432 | localhost | t              | t
>          |  888802358
>
I assume that you used the short install document from here:
http://postgres-xc.sourceforge.net/docs/1_0/install-short.html

I also imagine that you ran "SELECT pgxc_pool_reload();" after setting the
datanode on the Coordinator.
This looks definitely correct.

The problem I'm having is that after maybe 60 to 90 seconds or so after
> starting, the coordinator dies with a log message like
>
> FATAL:  Can not register Coordinator on GTM
>
This basically means that the Coordinator is trying to register to GTM with
a node name that is already used by another node.
What is the value of pgxc_node_name in each node's postgresql.conf?

The gtm is definitely running, I can see it in the process list.  The only
> changes to the default postgresql.conf's I've made is to specify the port
> of 15432 for datanode1, and (as part of an effort to fix this problem)
> uncomment the host ('localhost') and port (6666) of the gtm in all 3
> postgresql.conf files.
>
For a really basic setting, changing the port number of datanode1 is
enough, as well as specifying the port value with -p at startup.
The value specified by option -p has higher priority than the value of port
in postgresql.conf.


> Another thing I did to try to fix this was:
>
> alter node datanode1 with (primary = true, preferred = true);
>
Primary nodes are the nodes used in priority for replicated writes.
Preferred nodes are the nodes used in priority for replicated reads. So
changing this setting won't solve anything in your case.


>
> As you can see, that has affected the pgxc_node contents.  But that didn't
> have any impact on this crashing behavior.
>
> I tried starting using the method presented in the short version install
> to start everything, and I also tried starting up like this:
>
> gtm_ctl -Z gtm -D data_gtm -l gtm.log start
> sleep 1
> pg_ctl -Z coordinator -D data_coord1 -l coord1.log start
> sleep 1
> pg_ctl -Z datanode -D data_datanode1 -l datanode1.log start
>
> ..from the /usr/local/pgsql directory.  (I also tried specifying absolute
> paths.)
>
Absolute and non-absolute paths both work, like postgres.

In your case, I suggest to check for the values of pgxc_node_name in
postgresql.conf. Please be sure that they are set to the same values as
node_name in pgxc_node.
What you can also check is the data of the file called register.node in the
data folder of GTM. The data of this file is binary, but you may be able to
check the data folders of the nodes registered on GTM even by opening the
file manually.

Please also try the following thing if problem persists. This simple set of
commands allows to start a cluster with 1 Coordinator and 1 Datanode:
mkdir $HOME/pgsql/data_coord1
mkdir $HOME/pgsql/data_datanode1
mkdir $HOME/pgsql/data_gtm
initdb -D $HOME/pgsql/data_coord1 --nodename coord1
initdb -D $HOME/pgsql/data_datanode1 --nodename datanode1
initgtm -D $HOME/pgsql/data_gtm -Z gtm
gtm -D $HOME/pgsql/data_gtm &
postgres -X -p 15432 -D $HOME/pgsql/data_datanode1 &
postgres -C -D $HOME/pgsql/data_coord1 &
psql -c "CREATE NODE datanode1 WITH (TYPE = 'datanode', PORT = 15432)"
postgres
psql -c "SELECT pgxc_pool_reload()" postgres
createdb test
psql test

In my case, this works perfectly with XC 1.0.1, even with firewall and
SELinux in my box.

Then why not testing pgbench?
pgbench -i test
pgbench -c 24 -S test

Regards,
-- 
Michael Paquier
http://michael.otacoo.com