postgres-xc-general Mailing List for Postgres-XC (Page 43)

Well,  it seems that this is the way to connect it from the standard psql client.  But I'm getting there ;(

--------------
postgres-xc@adminuser-VirtualBox:~/coord$ whoami
postgres-xc
postgres-xc@adminuser-VirtualBox:~/coord$ echo $PSGSSLMODE
require
postgres-xc@adminuser-VirtualBox:~/coord$ echo $PGSSLCERT
/var/lib/postgres-xc/datanode1/server.crt
postgres-xc@adminuser-VirtualBox:~/coord$ ls -l server.*
-rw-rw-r-- 1 postgres-xc postgres-xc 4608 Feb 25 09:00 server.crt
-rw------- 1 postgres-xc postgres-xc 1679 Feb 25 09:00 server.key
-rw-rw-r-- 1 postgres-xc postgres-xc 3587 Feb 25 09:00 server.req
postgres-xc@adminuser-VirtualBox:~/coord$ psql -U postgres-xc -h localhost -p 5432
psql: certificate present, but not private key file "/var/lib/postgres-xc/.postgresql/postgresql.key"
postgres-xc@adminuser-VirtualBox:~/coord$ pwd
/var/lib/postgres-xc/coord
postgres-xc@adminuser-VirtualBox:~/coord$
--------------

Does the user 'postgres-xc' need to generate a private key file?

postgresql.key does not exist.  I take it that it IS NOT the same as server.key which is under /var/lib/postgres-xc/coord.  Am I correct?


From: Michael Paquier [mailto:mic...@gm...]
Sent: Monday, February 25, 2013 2:17 AM
To: Nikhil Sontakke
Cc: Maz Mohammadi; pos...@li...
Subject: Re: [Postgres-xc-general] can't access server through SSL


On Mon, Feb 25, 2013 at 4:10 PM, Nikhil Sontakke <ni...@st...<mailto:ni...@st...>> wrote:
Hi Maz,

> Is there a way to connect to the server using the psql by using a certificate?
>
Does the below work for you?

PSGSSLMODE=require PGSSLCERT=/path/to/agent.crt psql -d postgres -h
localhost -p 5432
Change PSGSSLMODE by PGSSLMODE. My 2c.
--
Michael

I downloaded the binaries, and it seems that it was built with openssl.  I'm not getting any errors about missing libraries or .so files.  

-maz

-----Original Message-----
From: Koichi Suzuki [mailto:koi...@gm...] 
Sent: Monday, February 25, 2013 6:07 AM
To: Maz Mohammadi
Cc: Nikhil Sontakke; pos...@li...
Subject: Re: [Postgres-xc-general] can't access server through SSL

Did you specify "--with-openssl" when you ran ./configure?   It is not
turned on by default.

Regards;
----------
Koichi Suzuki


2013/2/25 Michael Paquier <mic...@gm...>:
>
>
> On Mon, Feb 25, 2013 at 4:10 PM, Nikhil Sontakke <ni...@st...>
> wrote:
>>
>> Hi Maz,
>>
>> > Is there a way to connect to the server using the psql by using a 
>> > certificate?
>> >
>> Does the below work for you?
>>
>> PSGSSLMODE=require PGSSLCERT=/path/to/agent.crt psql -d postgres -h 
>> localhost -p 5432
>
> Change PSGSSLMODE by PGSSLMODE. My 2c.
> --
> Michael
>
> ----------------------------------------------------------------------
> -------- Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics Download AppDynamics Lite 
> for free today:
> http://p.sf.net/sfu/appdyn_d2d_feb
> _______________________________________________
> Postgres-xc-general mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>

Did you specify "--with-openssl" when you ran ./configure?   It is not
turned on by default.

Regards;
----------
Koichi Suzuki


2013/2/25 Michael Paquier <mic...@gm...>:
>
>
> On Mon, Feb 25, 2013 at 4:10 PM, Nikhil Sontakke <ni...@st...>
> wrote:
>>
>> Hi Maz,
>>
>> > Is there a way to connect to the server using the psql by using a
>> > certificate?
>> >
>> Does the below work for you?
>>
>> PSGSSLMODE=require PGSSLCERT=/path/to/agent.crt psql -d postgres -h
>> localhost -p 5432
>
> Change PSGSSLMODE by PGSSLMODE. My 2c.
> --
> Michael
>
> ------------------------------------------------------------------------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> http://p.sf.net/sfu/appdyn_d2d_feb
> _______________________________________________
> Postgres-xc-general mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>

On Mon, Feb 25, 2013 at 4:10 PM, Nikhil Sontakke <ni...@st...>wrote:

> Hi Maz,
>
> > Is there a way to connect to the server using the psql by using a
> certificate?
> >
> Does the below work for you?
>
> PSGSSLMODE=require PGSSLCERT=/path/to/agent.crt psql -d postgres -h
> localhost -p 5432
>
Change PSGSSLMODE by PGSSLMODE. My 2c.
-- 
Michael

Hi Maz,

> Is there a way to connect to the server using the psql by using a certificate?
>
Does the below work for you?

PSGSSLMODE=require PGSSLCERT=/path/to/agent.crt psql -d postgres -h
localhost -p 5432

Regards,
Nikhils

> So I hope this provides enough information and I'm in the right place.  Any help is greatly appreciated.  I've been twisting my head on this :(
>
> -maz
>
> ------------------------------------------------------------------------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> http://p.sf.net/sfu/appdyn_d2d_feb
> _______________________________________________
> Postgres-xc-general mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general



-- 
StormDB - http://www.stormdb.com
The Database Cloud
Postgres-XC Support and Service

On Mon, Feb 25, 2013 at 3:08 PM, kushal <kus...@gm...> wrote:

> I think it wont work for our requirements. Let me tell you the scenario.
>
> Suppose there is an existing PostgresSQL database instance DB1 with all
> types of unique constraints and foreign keys. So there is a PostgresSQL
> schema. Now it is required to just run the same schema file on XC
> coordinator without adding explicit distribute logic or modifying
> constraints. I think with current default distribution, it won't allow
> since there is restriction on number and nature of constraints.
>
Indeed. Constraints need to be enforceable to remote nodes in order to be
used. For example primary keys on distributed tables can only be used if
the distribution key contains the primary key.


> So as I see it, there can be two ways



> 1. Change the default distribution to replication or add distribute by
> replication somehow while creating tables.

Yeah there have been discussions in the past to change the default to
replication, however the conclusion was that we could not spot all the
potential problems especially with default set as replication.
You will need to hack the code a bit to achieve that. It is not that
difficult though.

> 2. Change the existing default distribution to allow multiple constraints
> or somehow control it at higher level. Can I just force the schema
> generation and then change the distribution to replication before inserting
> any data?
>
No, you can't enforce the schema generation directly before inserting data
as replication can only be defined with CREATE TABLE.

> What should be the way to go?
>
Number 1 clearly, hack the code and have the default distribution set as
REPLICATION if no DISTRIBUTE BY clause is specified in CREATE TABLE.
-- 
Michael

I think it wont work for our requirements. Let me tell you the scenario.

Suppose there is an existing PostgresSQL database instance DB1 with all
types of unique constraints and foreign keys. So there is a PostgresSQL
schema. Now it is required to just run the same schema file on XC
coordinator without adding explicit distribute logic or modifying
constraints. I think with current default distribution, it won't allow
since there is restriction on number and nature of constraints. So as I see
it, there can be two ways


   1. Change the default distribution to replication or add distribute by
   replication somehow while creating tables.
   2. Change the existing default distribution to allow multiple
   constraints or somehow control it at higher level. Can I just force the
   schema generation and then change the distribution to replication before
   inserting any data?

What should be the way to go?

--Kushal


On 25 February 2013 11:02, Michael Paquier <mic...@gm...>wrote:

>
>
> On Mon, Feb 25, 2013 at 2:12 PM, kushal <kus...@gm...> wrote:
>
>> Is there a way to change the default distribution to distribute by
>> replication during table generation?
>>
> No, there is no such option. You need to specify DISTRIBUTE BY
> REPLICATION. There were discussions in the XC MLs in the past about adding
> a GUC that would control default distribution type though. One of the main
> stopover of this feature is that you need to regenerate the DDL query on
> Coordinator to take care of cases where nodes have a different default
> distribution to be sure that things are consistent in the cluster. I hoped
> that event triggers could implement DDL deparsing in PG core, but it looks
> that this is not going in for 9.3, and some of the PG committers don't
> really like of being able to deparse DDL as it would mean that each new DDL
> command introduced would need to extend also the deparsing, adding
> unwelcome maintenance.
>
> Can I control the number of replicas of any table across datanodes and
>> also control on what set of datanodes, replica/s can be generated?
>>
> Yes. You can control where data of tables is replicated.
> For example, let's imagine that you have a cluster with Datanodes dn1,
> dn2, dn3 and dn4, you can create a table in such a way that its data is
> only replicated on nodes dn2 and dn4 with that:
> CREATE TABLE aa (a int) DISTRIBUTE BY REPLICATION TO NODE dn2,dn4;
>
> You can also specify a group of nodes:
> CREATE NODE GROUP mygroup WITH (dn2,dn4);
> CREATE TABLE aa (a int) DISTRIBUTE BY REPLICATION TO GROUP mygroup;
> --
> Michael
>

HI Maz,
Quickly, I can say, that the same connection methods as Postgres will work
with Postgres-XC. So, check if you can connect to a PostgreSQL server first
and then try XC.

In general XC configuration is more complicated than PostgreSQL, so it will
help, if you try things on Postgres before using Postgres-XC.

On Mon, Feb 25, 2013 at 10:42 AM, Maz Mohammadi <mmo...@pe...>wrote:

> Hi everyone,
>
> I posted this question on postgres-general, but found out that this is a
> better spot for it.  So let me start again.
>
> I need to verify how I can connect to a postgres-xc(which I know is not
> the same as postgres) server from a JDBC datasource on tomcat using
> certificate authentication.  I'm new to postgres, and I installed it
> (perhaps by mistake postgres-xc) on my Ubuntu virtual box.  As I've
> understood it, there are many moving parts, and I figured out that I need
> to keep the configuration files in sync.  So under /var/lib/postgres-xc
> there are these folders.
> coord,
> datanode1, and
> datanode2
>
> so far I had to start all of them + gtm (global transaction manager).
>  Then I can connect to the server via psql
>  postgres-xc@adminuser-VirtualBox:~/datanode2$ psql -h localhost
>  psql: FATAL:  connection requires a valid client certificate
>  FATAL:  no pg_hba.conf entry for host "127.0.0.1", user "postgres-xc",
>  database "postgres", SSL off
>  postgres-xc@adminuser-VirtualBox:~/datanode2$
>
> I have the following lines in the pg_hba.conf file under
> /var/lib/postgres-xc/coord
> --------
> local all postgres-xc peer
> hostssl all             all             127.0.0.1/32            cert
> --------
> And ssl=on has been set in postgresql.conf for all nodes (datanode1,2 and
> coordinator).
>
> In all directories, I've placed server.key, server.crt and copied
> server.crt into root.crt.
>
> The client application is a tomcat app server, where I'm trying to create
> a JDBC data source.  I'm using JDBC3 drivers.  My jdbc url is
> jdbc:postgresql://localhost:5432:testdb?ssl=true........5432 is the
> coordinator's port.  And when I try to test it, I get the following:
> "Connection attempt failed: FATAL: connection requires a valid client
> certificate"
> I get the same result when I try port 15432, which is the listen port of
> the one of the data nodes.
>
> As for the documentation on
> http://www.postgresql.org/docs/9.2/interactive/ssl-tcp.html, I've already
> been through it, and created the certificates and put them in java's
> keystore file.
>  -Djavax.net.ssl.trustStore=/home/adminuser/pentaho/keycerts/mazstore
>  -Djavax.net.ssl.trustStorePassword=password
>
> Is there a way to connect to the server using the psql by using a
> certificate?
>
> So I hope this provides enough information and I'm in the right place.
>  Any help is greatly appreciated.  I've been twisting my head on this :(
>
> -maz
>
>
> ------------------------------------------------------------------------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> http://p.sf.net/sfu/appdyn_d2d_feb
> _______________________________________________
> Postgres-xc-general mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>



-- 
Best Wishes,
Ashutosh Bapat
EntepriseDB Corporation
The Enterprise Postgres Company

On Mon, Feb 25, 2013 at 2:12 PM, kushal <kus...@gm...> wrote:

> Is there a way to change the default distribution to distribute by
> replication during table generation?
>
No, there is no such option. You need to specify DISTRIBUTE BY REPLICATION.
There were discussions in the XC MLs in the past about adding a GUC that
would control default distribution type though. One of the main stopover of
this feature is that you need to regenerate the DDL query on Coordinator to
take care of cases where nodes have a different default distribution to be
sure that things are consistent in the cluster. I hoped that event triggers
could implement DDL deparsing in PG core, but it looks that this is not
going in for 9.3, and some of the PG committers don't really like of being
able to deparse DDL as it would mean that each new DDL command introduced
would need to extend also the deparsing, adding unwelcome maintenance.

Can I control the number of replicas of any table across datanodes and also
> control on what set of datanodes, replica/s can be generated?
>
Yes. You can control where data of tables is replicated.
For example, let's imagine that you have a cluster with Datanodes dn1, dn2,
dn3 and dn4, you can create a table in such a way that its data is only
replicated on nodes dn2 and dn4 with that:
CREATE TABLE aa (a int) DISTRIBUTE BY REPLICATION TO NODE dn2,dn4;

You can also specify a group of nodes:
CREATE NODE GROUP mygroup WITH (dn2,dn4);
CREATE TABLE aa (a int) DISTRIBUTE BY REPLICATION TO GROUP mygroup;
-- 
Michael

Hi

Is there a way to change the default distribution to distribute by
replication during table generation?

Can I control the number of replicas of any table across datanodes and also
control on what set of datanodes, replica/s can be generated?

Thanks
Kushal

Hi everyone,

I posted this question on postgres-general, but found out that this is a better spot for it.  So let me start again.

I need to verify how I can connect to a postgres-xc(which I know is not the same as postgres) server from a JDBC datasource on tomcat using certificate authentication.  I'm new to postgres, and I installed it (perhaps by mistake postgres-xc) on my Ubuntu virtual box.  As I've understood it, there are many moving parts, and I figured out that I need to keep the configuration files in sync.  So under /var/lib/postgres-xc there are these folders.
coord,
datanode1, and 
datanode2

so far I had to start all of them + gtm (global transaction manager).  Then I can connect to the server via psql
 postgres-xc@adminuser-VirtualBox:~/datanode2$ psql -h localhost
 psql: FATAL:  connection requires a valid client certificate
 FATAL:  no pg_hba.conf entry for host "127.0.0.1", user "postgres-xc", 
 database "postgres", SSL off 
 postgres-xc@adminuser-VirtualBox:~/datanode2$

I have the following lines in the pg_hba.conf file under /var/lib/postgres-xc/coord
--------
local all postgres-xc peer
hostssl all             all             127.0.0.1/32            cert
--------
And ssl=on has been set in postgresql.conf for all nodes (datanode1,2 and coordinator).

In all directories, I've placed server.key, server.crt and copied server.crt into root.crt.	

The client application is a tomcat app server, where I'm trying to create a JDBC data source.  I'm using JDBC3 drivers.  My jdbc url is jdbc:postgresql://localhost:5432:testdb?ssl=true........5432 is the coordinator's port.  And when I try to test it, I get the following:
"Connection attempt failed: FATAL: connection requires a valid client certificate"
I get the same result when I try port 15432, which is the listen port of the one of the data nodes.

As for the documentation on http://www.postgresql.org/docs/9.2/interactive/ssl-tcp.html, I've already been through it, and created the certificates and put them in java's keystore file.
 -Djavax.net.ssl.trustStore=/home/adminuser/pentaho/keycerts/mazstore 
 -Djavax.net.ssl.trustStorePassword=password

Is there a way to connect to the server using the psql by using a certificate?  

So I hope this provides enough information and I'm in the right place.  Any help is greatly appreciated.  I've been twisting my head on this :(

-maz

Thanks Michael for the advice.

For more tips you should do, please look into a bash script pgxc_ctl
available at
https://github.com/koichi-szk/PGXC-Tools/tree/master/pgxc_ctl

Regards;
----------
Koichi Suzuki


2013/2/22 Michael Paquier <mic...@gm...>:
>
>
> On Fri, Feb 22, 2013 at 2:29 PM, kushal <kus...@gm...> wrote:
>>
>> Hi
>>
>> I am trying to create a setup with 2 coordinators and 2 datanodes on one
>> server. I am able to get the gtm and nodes up and running.
>>
>> Ports:
>> Data Node 1: 15442
>> Data Node 2: 15443
>> Coordinator 1: 15432
>> Coordinator 1: 15433
>>
>> Next I did 'psql -p 15432 postgres' and created two datanodes
>> CREATE NODE dn1 WITH (TYPE='datanode', PORT=15442);
>> CREATE NODE dn2 WITH (TYPE='datanode', PORT=15443);
>>
>> After that I create one database: 'Create database mydb'
>>
>> Now I can see mydb database on datanodes 1 and 2 and coordinator 1.
>> Also I can see the datanode 1 and 2 when I execute 'select * from
>> pgxc_node' on coordinator 1.
>>
>> But mydb database does not exist and datanodes 1 and 2 are not present in
>> pgxc_node table on coordinator 2.
>>
>> Is there anything I have missed?
>
> Yes, 2 things:
> - You have 2 Coordinators, so you need also to register Coordinator 2 on
> Coordinator 1 like that:
> psql -p 15432 -c "CREATE NODE coord2 WITH (TYPE='coordinator', PORT=15433)"
> And register Coordinator 1 on Coordinator 2 like that:
> psql -p 15433 -c "CREATE NODE coord1 WITH (TYPE='coordinator', PORT=15432)"
> - 2nd thing you forgot is to run this command on each Coordinator:
> select pgxc_pool_reload();
> This updates the pooler cache located on each Coordinator with latest node
> information.
>
>> I think this might be a very basic thing and I am sorry to bother you
>> guys. I tried and failed to search through the xc mailing lists for older
>> posts on similar topic. Is there a way to quickly filter the older posts for
>> quick check?
>
> Everything's here:
> http://sourceforge.net/search/?group_id=311227&type_of_search=mlists&source=navbar
> --
> Michael
>
> ------------------------------------------------------------------------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> http://p.sf.net/sfu/appdyn_d2d_feb
> _______________________________________________
> Postgres-xc-general mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>

Thanks Michael. It worked.

On 22 February 2013 12:35, Michael Paquier <mic...@gm...>wrote:

>
>
> On Fri, Feb 22, 2013 at 3:45 PM, kushal <kus...@gm...> wrote:
>
>> Its still not working. So here is what I have done. I dropped mydb
>> database. I dropped data nodes from both coordinators.
>> Then I ran:
>> psql -p 15432 -c "CREATE NODE coord2 WITH (TYPE='coordinator',
>> PORT=15433)"
>>  psql -p 15433 -c "CREATE NODE coord1 WITH (TYPE='coordinator',
>> PORT=15432)"
>> select pgxc_pool_reload(); on each coordinator
>>
>> Then I created datanode 1 on coord1.
>> Again I ran select pgxc_pool_reload(); on each coordinator. Still I
>> cannot see datanode1 entry on coordinator 2.
>>
> Meh? I don't really think you understood what I was explaining. You need
> to create *all* the remote nodes on each Coordinator.
> So on Coordinator 1: create Co2, Dn1 and Dn2.
> and on Coordinator 2: create Co1, Dn1 and Dn2.
>
> So...
> Here is what you need to launch as commands on your server:
> # Register Coordinators on each other Coordinator
>
> psql -p 15432 -c "CREATE NODE coord2 WITH (TYPE='coordinator', PORT=15433)"
>  psql -p 15433 -c "CREATE NODE coord1 WITH (TYPE='coordinator',
> PORT=15432)"
> # Register Datanodes on Coordinator 1 and update cache
> psql -p 15432 -c "CREATE NODE dn1 WITH (TYPE='datanode', PORT=15442)"
> psql -p 15432 -c "CREATE NODE dn2 WITH (TYPE='datanode', PORT=15443)"
> psql -p 15432 -c "select pgxc_pool_reload()"
> # Register Datanodes on Coordinator 2 and update cache
> psql -p 15433 -c "CREATE NODE dn1 WITH (TYPE='datanode', PORT=15442)"
> psql -p 15433 -c "CREATE NODE dn2 WITH (TYPE='datanode', PORT=15443)"
> psql -p 15433 -c "select pgxc_pool_reload()"
>
> And you are done.
>
>
>> Also I noticed that under the table pgxc_node on coord1, there is an
>> entry for coord1 with port as 5432, whereas the actual port for coord1 is
>> 15432.
>>
> Don't worry about that this is a dummy entry.
>
>
>> Similarly on coord2, the self port under pgxc_node is 5432, but actual
>> value is 15433. I am not sure whether this is ok.
>>
> This is perfectly fine. You could still use ALTER NODE to update the port
> value to your needs, but this is not necessary at all as a node is not
> going to connect to itself
> --
> Michael
>

On Fri, Feb 22, 2013 at 3:45 PM, kushal <kus...@gm...> wrote:

> Its still not working. So here is what I have done. I dropped mydb
> database. I dropped data nodes from both coordinators.
> Then I ran:
> psql -p 15432 -c "CREATE NODE coord2 WITH (TYPE='coordinator', PORT=15433)"
> psql -p 15433 -c "CREATE NODE coord1 WITH (TYPE='coordinator', PORT=15432)"
> select pgxc_pool_reload(); on each coordinator
>
> Then I created datanode 1 on coord1.
> Again I ran select pgxc_pool_reload(); on each coordinator. Still I cannot
> see datanode1 entry on coordinator 2.
>
Meh? I don't really think you understood what I was explaining. You need to
create *all* the remote nodes on each Coordinator.
So on Coordinator 1: create Co2, Dn1 and Dn2.
and on Coordinator 2: create Co1, Dn1 and Dn2.

So...
Here is what you need to launch as commands on your server:
# Register Coordinators on each other Coordinator
psql -p 15432 -c "CREATE NODE coord2 WITH (TYPE='coordinator', PORT=15433)"
psql -p 15433 -c "CREATE NODE coord1 WITH (TYPE='coordinator', PORT=15432)"
# Register Datanodes on Coordinator 1 and update cache
psql -p 15432 -c "CREATE NODE dn1 WITH (TYPE='datanode', PORT=15442)"
psql -p 15432 -c "CREATE NODE dn2 WITH (TYPE='datanode', PORT=15443)"
psql -p 15432 -c "select pgxc_pool_reload()"
# Register Datanodes on Coordinator 2 and update cache
psql -p 15433 -c "CREATE NODE dn1 WITH (TYPE='datanode', PORT=15442)"
psql -p 15433 -c "CREATE NODE dn2 WITH (TYPE='datanode', PORT=15443)"
psql -p 15433 -c "select pgxc_pool_reload()"

And you are done.


> Also I noticed that under the table pgxc_node on coord1, there is an entry
> for coord1 with port as 5432, whereas the actual port for coord1 is 15432.
>
Don't worry about that this is a dummy entry.


> Similarly on coord2, the self port under pgxc_node is 5432, but actual
> value is 15433. I am not sure whether this is ok.
>
This is perfectly fine. You could still use ALTER NODE to update the port
value to your needs, but this is not necessary at all as a node is not
going to connect to itself
-- 
Michael

Its still not working. So here is what I have done. I dropped mydb
database. I dropped data nodes from both coordinators.

Then I ran:
psql -p 15432 -c "CREATE NODE coord2 WITH (TYPE='coordinator', PORT=15433)"
psql -p 15433 -c "CREATE NODE coord1 WITH (TYPE='coordinator', PORT=15432)"
select pgxc_pool_reload(); on each coordinator

Then I created datanode 1 on coord1.
Again I ran select pgxc_pool_reload(); on each coordinator. Still I cannot
see datanode1 entry on coordinator 2.

Also I noticed that under the table pgxc_node on coord1, there is an entry
for coord1 with port as 5432, whereas the actual port for coord1 is 15432.
Similarly on coord2, the self port under pgxc_node is 5432, but actual
value is 15433. I am not sure whether this is ok.

Regards
Kushal


On 22 February 2013 11:37, Michael Paquier <mic...@gm...>wrote:

>
>
> On Fri, Feb 22, 2013 at 2:29 PM, kushal <kus...@gm...> wrote:
>
>> Hi
>>
>> I am trying to create a setup with 2 coordinators and 2 datanodes on one
>> server. I am able to get the gtm and nodes up and running.
>>
>> Ports:
>> Data Node 1: 15442
>> Data Node 2: 15443
>> Coordinator 1: 15432
>> Coordinator 1: 15433
>>
>> Next I did 'psql -p 15432 postgres' and created two datanodes
>> CREATE NODE dn1 WITH (TYPE='datanode', PORT=15442);
>> CREATE NODE dn2 WITH (TYPE='datanode', PORT=15443);
>>
>> After that I create one database: 'Create database mydb'
>>
>> Now I can see mydb database on datanodes 1 and 2 and coordinator 1.
>> Also I can see the datanode 1 and 2 when I execute 'select * from
>> pgxc_node' on coordinator 1.
>>
>> But mydb database does not exist and datanodes 1 and 2 are not present in
>> pgxc_node table on coordinator 2.
>>
>> Is there anything I have missed?
>>
> Yes, 2 things:
> - You have 2 Coordinators, so you need also to register Coordinator 2 on
> Coordinator 1 like that:
> psql -p 15432 -c "CREATE NODE coord2 WITH (TYPE='coordinator', PORT=15433)"
> And register Coordinator 1 on Coordinator 2 like that:
> psql -p 15433 -c "CREATE NODE coord1 WITH (TYPE='coordinator', PORT=15432)"
> - 2nd thing you forgot is to run this command on each Coordinator:
> select pgxc_pool_reload();
> This updates the pooler cache located on each Coordinator with latest node
> information.
>
> I think this might be a very basic thing and I am sorry to bother you
>> guys. I tried and failed to search through the xc mailing lists for older
>> posts on similar topic. Is there a way to quickly filter the older posts
>> for quick check?
>>
> Everything's here:
>
> http://sourceforge.net/search/?group_id=311227&type_of_search=mlists&source=navbar
> --
> Michael
>

On Fri, Feb 22, 2013 at 2:29 PM, kushal <kus...@gm...> wrote:

> Hi
>
> I am trying to create a setup with 2 coordinators and 2 datanodes on one
> server. I am able to get the gtm and nodes up and running.
>
> Ports:
> Data Node 1: 15442
> Data Node 2: 15443
> Coordinator 1: 15432
> Coordinator 1: 15433
>
> Next I did 'psql -p 15432 postgres' and created two datanodes
> CREATE NODE dn1 WITH (TYPE='datanode', PORT=15442);
> CREATE NODE dn2 WITH (TYPE='datanode', PORT=15443);
>
> After that I create one database: 'Create database mydb'
>
> Now I can see mydb database on datanodes 1 and 2 and coordinator 1.
> Also I can see the datanode 1 and 2 when I execute 'select * from
> pgxc_node' on coordinator 1.
>
> But mydb database does not exist and datanodes 1 and 2 are not present in
> pgxc_node table on coordinator 2.
>
> Is there anything I have missed?
>
Yes, 2 things:
- You have 2 Coordinators, so you need also to register Coordinator 2 on
Coordinator 1 like that:
psql -p 15432 -c "CREATE NODE coord2 WITH (TYPE='coordinator', PORT=15433)"
And register Coordinator 1 on Coordinator 2 like that:
psql -p 15433 -c "CREATE NODE coord1 WITH (TYPE='coordinator', PORT=15432)"
- 2nd thing you forgot is to run this command on each Coordinator:
select pgxc_pool_reload();
This updates the pooler cache located on each Coordinator with latest node
information.

I think this might be a very basic thing and I am sorry to bother you guys.
> I tried and failed to search through the xc mailing lists for older posts
> on similar topic. Is there a way to quickly filter the older posts for
> quick check?
>
Everything's here:
http://sourceforge.net/search/?group_id=311227&type_of_search=mlists&source=navbar
-- 
Michael

Hi

I am trying to create a setup with 2 coordinators and 2 datanodes on one
server. I am able to get the gtm and nodes up and running.

Ports:
Data Node 1: 15442
Data Node 2: 15443
Coordinator 1: 15432
Coordinator 1: 15433

Next I did 'psql -p 15432 postgres' and created two datanodes
CREATE NODE dn1 WITH (TYPE='datanode', PORT=15442);
CREATE NODE dn2 WITH (TYPE='datanode', PORT=15443);

After that I create one database: 'Create database mydb'

Now I can see mydb database on datanodes 1 and 2 and coordinator 1.
Also I can see the datanode 1 and 2 when I execute 'select * from
pgxc_node' on coordinator 1.

But mydb database does not exist and datanodes 1 and 2 are not present in
pgxc_node table on coordinator 2.

Is there anything I have missed?


I think this might be a very basic thing and I am sorry to bother you guys.
I tried and failed to search through the xc mailing lists for older posts
on similar topic. Is there a way to quickly filter the older posts for
quick check?

Thanks
Kushal

Thanks guys. Issue is resolved.


On 21 February 2013 04:04, Michael Paquier <mic...@gm...>wrote:

>
>
> On Thu, Feb 21, 2013 at 8:47 PM, kushal <kus...@gm...> wrote:
>
>> LOG:  database system was interrupted; last known up at 2013-02-21
>> 02:50:33 PST
>> FATAL:  lock file "/tmp/.s.PGPOOL.6667.lock" already exists
>> HINT:  Is another postmaster (PID 12892) using socket file
>> "/tmp/.s.PGPOOL.6667"?
>> LOG:  database system was not properly shut down; automatic recovery in
>> progress
>> LOG:  pool manager process (PID 12907) exited with exit code 1
>> LOG:  terminating any other active server processes
>> LOG:  startup process (PID 12906) exited with exit code 2
>> LOG:  aborting startup due to startup process failure
>>
>>
>> Can anyone please help?
>>
> Simply set pooler_port to different values in each Coordinator's
> postgresql.conf. As 6667 is the default value, this error is encountered
> each time a second Coordiantor comes up.
> --
> Michael
>

On Thu, Feb 21, 2013 at 8:47 PM, kushal <kus...@gm...> wrote:

> LOG:  database system was interrupted; last known up at 2013-02-21
> 02:50:33 PST
> FATAL:  lock file "/tmp/.s.PGPOOL.6667.lock" already exists
> HINT:  Is another postmaster (PID 12892) using socket file
> "/tmp/.s.PGPOOL.6667"?
> LOG:  database system was not properly shut down; automatic recovery in
> progress
> LOG:  pool manager process (PID 12907) exited with exit code 1
> LOG:  terminating any other active server processes
> LOG:  startup process (PID 12906) exited with exit code 2
> LOG:  aborting startup due to startup process failure
>
>
> Can anyone please help?
>
Simply set pooler_port to different values in each Coordinator's
postgresql.conf. As 6667 is the default value, this error is encountered
each time a second Coordiantor comes up.
-- 
Michael

Hi Kushal,

Please ensure that the pooler_port value is different for both the coordinators.

Regards,
Nikhils

On Thu, Feb 21, 2013 at 5:29 PM, Mason Sharp <ma...@st...> wrote:
>
>
> On Thu, Feb 21, 2013 at 6:49 AM, kushal <kus...@gm...> wrote:
>>
>> Forgot to add. This whole setup is tried on one single server.
>>
>>
>> On 21 February 2013 03:47, kushal <kus...@gm...> wrote:
>>>
>>> Hi
>>>
>>> I am using XC 1.0.2. I am trying to create the whole setup including gtm,
>>> 2 gtm proxies, 2 coordinators and 2 datanodes.
>>>
>>> Ports Configured:
>>> GTM: 16666
>>> GTM Proxy 1: 16667
>>> GTM Proxy 1: 16668
>>> Data Node 1: 15442
>>> Data Node 2: 15443
>>> Coordinator 1: 15432
>>> Coordinator 1: 15433
>>>
>>> Coordinator 1 has gtm port configured to 16667 and Coordinator 2 to
>>> 16668.
>>>
>>> Next I try to start all the nodes.
>>> gtm_ctl start -Z gtm -D /var/lib/pgsql/gtm
>>> gtm_ctl start -Z gtm_proxy -D /var/lib/pgsql/gtm_proxy1
>>> gtm_ctl start -Z gtm_proxy -D /var/lib/pgsql/gtm_proxy2
>>> pg_ctl start -Z datanode -D /var/lib/pgsql/data1
>>> pg_ctl start -Z datanode -D /var/lib/pgsql/data2
>>> pg_ctl start -Z coordinator -D /var/lib/pgsql/coord1
>>> pg_ctl start -Z coordinator -D /var/lib/pgsql/coord2
>>>
>>> Everything works fine till coord1. In the last step while starting
>>> coord2, I get this error:
>>>
>>> LOG:  database system was interrupted; last known up at 2013-02-21
>>> 02:50:33 PST
>>> FATAL:  lock file "/tmp/.s.PGPOOL.6667.lock" already exists
>
>
> Are you using the same port number in the coordinators for each one's
> connection pool?
>
>>>
>>> HINT:  Is another postmaster (PID 12892) using socket file
>>> "/tmp/.s.PGPOOL.6667"?
>>> LOG:  database system was not properly shut down; automatic recovery in
>>> progress
>>> LOG:  pool manager process (PID 12907) exited with exit code 1
>>> LOG:  terminating any other active server processes
>>> LOG:  startup process (PID 12906) exited with exit code 2
>>> LOG:  aborting startup due to startup process failure
>>>
>>>
>>> Can anyone please help?
>>>
>>>
>>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Everyone hates slow websites. So do we.
>> Make your web apps faster with AppDynamics
>> Download AppDynamics Lite for free today:
>> http://p.sf.net/sfu/appdyn_d2d_feb
>> _______________________________________________
>> Postgres-xc-general mailing list
>> Pos...@li...
>> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>>
>
>
>
> --
> Mason Sharp
>
> StormDB - http://www.stormdb.com
> The Database Cloud
> Postgres-XC Support and Services
> ------------------------------------------------------------------------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> http://p.sf.net/sfu/appdyn_d2d_feb
> _______________________________________________
> Postgres-xc-general mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>



-- 
StormDB - http://www.stormdb.com
The Database Cloud
Postgres-XC Support and Service

On Thu, Feb 21, 2013 at 6:49 AM, kushal <kus...@gm...> wrote:

> Forgot to add. This whole setup is tried on one single server.
>
>
> On 21 February 2013 03:47, kushal <kus...@gm...> wrote:
>
>> Hi
>>
>> I am using XC 1.0.2. I am trying to create the whole setup including gtm,
>> 2 gtm proxies, 2 coordinators and 2 datanodes.
>>
>> Ports Configured:
>> GTM: 16666
>> GTM Proxy 1: 16667
>> GTM Proxy 1: 16668
>> Data Node 1: 15442
>> Data Node 2: 15443
>> Coordinator 1: 15432
>> Coordinator 1: 15433
>>
>> Coordinator 1 has gtm port configured to 16667 and Coordinator 2 to 16668.
>>
>> Next I try to start all the nodes.
>> gtm_ctl start -Z gtm -D /var/lib/pgsql/gtm
>> gtm_ctl start -Z gtm_proxy -D /var/lib/pgsql/gtm_proxy1
>> gtm_ctl start -Z gtm_proxy -D /var/lib/pgsql/gtm_proxy2
>> pg_ctl start -Z datanode -D /var/lib/pgsql/data1
>> pg_ctl start -Z datanode -D /var/lib/pgsql/data2
>> pg_ctl start -Z coordinator -D /var/lib/pgsql/coord1
>> pg_ctl start -Z coordinator -D /var/lib/pgsql/coord2
>>
>> Everything works fine till coord1. In the last step while starting
>> coord2, I get this error:
>>
>> LOG:  database system was interrupted; last known up at 2013-02-21
>> 02:50:33 PST
>> FATAL:  lock file "/tmp/.s.PGPOOL.6667.lock" already exists
>>
>
Are you using the same port number in the coordinators for each one's
connection pool?


> HINT:  Is another postmaster (PID 12892) using socket file
>> "/tmp/.s.PGPOOL.6667"?
>> LOG:  database system was not properly shut down; automatic recovery in
>> progress
>> LOG:  pool manager process (PID 12907) exited with exit code 1
>> LOG:  terminating any other active server processes
>> LOG:  startup process (PID 12906) exited with exit code 2
>> LOG:  aborting startup due to startup process failure
>>
>>
>> Can anyone please help?
>>
>>
>>
>>
>
>
> ------------------------------------------------------------------------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> http://p.sf.net/sfu/appdyn_d2d_feb
> _______________________________________________
> Postgres-xc-general mailing list
> Pos...@li...
> https://lists.sourceforge.net/lists/listinfo/postgres-xc-general
>
>


-- 
Mason Sharp

StormDB - http://www.stormdb.com
The Database Cloud
Postgres-XC Support and Services

Forgot to add. This whole setup is tried on one single server.

On 21 February 2013 03:47, kushal <kus...@gm...> wrote:

> Hi
>
> I am using XC 1.0.2. I am trying to create the whole setup including gtm,
> 2 gtm proxies, 2 coordinators and 2 datanodes.
>
> Ports Configured:
> GTM: 16666
> GTM Proxy 1: 16667
> GTM Proxy 1: 16668
> Data Node 1: 15442
> Data Node 2: 15443
> Coordinator 1: 15432
> Coordinator 1: 15433
>
> Coordinator 1 has gtm port configured to 16667 and Coordinator 2 to 16668.
>
> Next I try to start all the nodes.
> gtm_ctl start -Z gtm -D /var/lib/pgsql/gtm
> gtm_ctl start -Z gtm_proxy -D /var/lib/pgsql/gtm_proxy1
> gtm_ctl start -Z gtm_proxy -D /var/lib/pgsql/gtm_proxy2
> pg_ctl start -Z datanode -D /var/lib/pgsql/data1
> pg_ctl start -Z datanode -D /var/lib/pgsql/data2
> pg_ctl start -Z coordinator -D /var/lib/pgsql/coord1
> pg_ctl start -Z coordinator -D /var/lib/pgsql/coord2
>
> Everything works fine till coord1. In the last step while starting coord2,
> I get this error:
>
> LOG:  database system was interrupted; last known up at 2013-02-21
> 02:50:33 PST
> FATAL:  lock file "/tmp/.s.PGPOOL.6667.lock" already exists
> HINT:  Is another postmaster (PID 12892) using socket file
> "/tmp/.s.PGPOOL.6667"?
> LOG:  database system was not properly shut down; automatic recovery in
> progress
> LOG:  pool manager process (PID 12907) exited with exit code 1
> LOG:  terminating any other active server processes
> LOG:  startup process (PID 12906) exited with exit code 2
> LOG:  aborting startup due to startup process failure
>
>
> Can anyone please help?
>
>
>
>

Hi

I am using XC 1.0.2. I am trying to create the whole setup including gtm, 2
gtm proxies, 2 coordinators and 2 datanodes.

Ports Configured:
GTM: 16666
GTM Proxy 1: 16667
GTM Proxy 1: 16668
Data Node 1: 15442
Data Node 2: 15443
Coordinator 1: 15432
Coordinator 1: 15433

Coordinator 1 has gtm port configured to 16667 and Coordinator 2 to 16668.

Next I try to start all the nodes.
gtm_ctl start -Z gtm -D /var/lib/pgsql/gtm
gtm_ctl start -Z gtm_proxy -D /var/lib/pgsql/gtm_proxy1
gtm_ctl start -Z gtm_proxy -D /var/lib/pgsql/gtm_proxy2
pg_ctl start -Z datanode -D /var/lib/pgsql/data1
pg_ctl start -Z datanode -D /var/lib/pgsql/data2
pg_ctl start -Z coordinator -D /var/lib/pgsql/coord1
pg_ctl start -Z coordinator -D /var/lib/pgsql/coord2

Everything works fine till coord1. In the last step while starting coord2,
I get this error:

LOG:  database system was interrupted; last known up at 2013-02-21 02:50:33
PST
FATAL:  lock file "/tmp/.s.PGPOOL.6667.lock" already exists
HINT:  Is another postmaster (PID 12892) using socket file
"/tmp/.s.PGPOOL.6667"?
LOG:  database system was not properly shut down; automatic recovery in
progress
LOG:  pool manager process (PID 12907) exited with exit code 1
LOG:  terminating any other active server processes
LOG:  startup process (PID 12906) exited with exit code 2
LOG:  aborting startup due to startup process failure


Can anyone please help?

Mr. Ashutosh Bapat,

Thank you for your fast response & confirmation that FQS is executed in parallel. I was going to send you screenshots of the evolution of the query, but it did run in parallel perfectly. Earlier when I ran the query it ran "node by node" from the last node to the first node. Is there any logic that would cause this behavior - perhaps I need rest :.)

Best regards,

Arni

From: Ashutosh Bapat [mailto:ash...@en...]
Sent: Tuesday, February 19, 2013 11:08 PM
To: Arni Sumarlidason
Cc: pos...@li...
Subject: Re: [Postgres-xc-general] FQS/RGQ Optimizations

Hi Arni,
What do you mean by "in the execution of the FQS logic -- it seems to query one node at a time instead of in parallel"? What gave you that impression? When a query is fired on multiple nodes, it's done in-parallel or asynchronously.

For better understanding of how query is processed in Postgres-XC, please watch following video -
http://www.youtube.com/watch?v=g_9CYat8nkY<https://console.mxlogic.com/redir/?2eosod7bNEVd79JYsOMOyqejhOr015GJPr8vO-6dDoCicD4hkQsyg7SrQ2E7dTVeZXTLuZXCXCZEjrDmxfy6Hrc5v2vNmzaDE4endL8IcCQmn4mmmmrCzASMedwLQzh0qmMJmAOcvEn3X5OfM-eDCBQQg2U3z60I90eq80MJeiIxjpOH2xKvxYY1NJ4SOrpop73zhO-Urgf0s>
On Wed, Feb 20, 2013 at 8:01 AM, Arni Sumarlidason <Arn...@md...<mailto:Arn...@md...>> wrote:
Good Evening all,

Thank you for your support over the last few days, really appreciate it. :)

I have a question regarding query optimization, I have one table distributed by HASH, and another by REPLICATE. When attempting to compare values between the two tables[1] I noticed that in the execution of the FQS logic -- it seems to query one node at a time instead of in parallel. I google'd FQS and found an article by Michael Paquier and played with disabling the feature. However disabling it changed the execution method to _REMOTE_TABLE_QUERY_ -- this also did not seem to be parallel. Is there any way to optimize for __REMOTE_GROUP_QUERY__?

Thank you again,

Arni Sumarlidason

[1] SELECT * FROM data m, world_boundaries b WHERE ((b.cntry_name ILIKE '%Iceland%')) AND m.point_geom && b.wkb_geometry AND st_intersects(m.point_geom, b.wkb_geometry);
Data - Hash'd
world_boundaries - Replicated


------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_feb<https://console.mxlogic.com/redir/?4sMUMqenzhOqejrUVBxB4QsCzAS03mWzL0TaeJVGOO1q0GwCwaK-gdTVeZXTLuZXCXCZEjrDmxfy6Hrc5v2vNmzaDE4endL8IcCQmn4mmmmrCzASMedwLQzh0qmMJmAOcvEn3X5OfM-eDCBQQg2U3z60I90eq80MJeiIxjpOH2xKvxYY1NJcSOrpop73zhO-UrWlX5>
_______________________________________________
Postgres-xc-general mailing list
Pos...@li...<mailto:Pos...@li...>
https://lists.sourceforge.net/lists/listinfo/postgres-xc-general<https://console.mxlogic.com/redir/?zC763hOYqejhOrv7cIcECzAQsCOsVHkiP3VkDjGlrzrTg0fY3sfBitcfBisEeRNqJ9Ao_gK7SbAvxYtfdTVeZXTLuZXCXCZEjrDmxfy6Hrc5v2vNmzaDE4endL8IcCQmn4mmmmrCzASMedwLQzh0qmMJmAOcvEn3X5OfM-eDCBQQg2U3z60I90eq80MJeiIxjpOH2xKvxYY1NJASOrpop73zhO-Ur8tlHf_Fa0K-C>



--
Best Wishes,
Ashutosh Bapat
EntepriseDB Corporation
The Enterprise Postgres Company