Re: [GENERAL] PGPool in Master-Master mode, is it possible? - Mailing list pgsql-general

From Tatsuo Ishii
Subject Re: [GENERAL] PGPool in Master-Master mode, is it possible?
Date
Msg-id 20171116.080312.1344652738002914432.t-ishii@sraoss.co.jp
Whole thread Raw
In response to Re: [GENERAL] PGPool in Master-Master mode, is it possible?  (Tatsuo Ishii <ishii@sraoss.co.jp>)
List pgsql-general
I am not familiar with swarm cluster nor repmgr. So I wonder why you
wouldn't be able to get to the goal you are trying to achieve without
repmgr. Can you please elaborate more?

Best regards,
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese:http://www.sraoss.co.jp

> I am also interested in this set-up
> My idea is to run pgpool in a docker swarm cluster and to have two instances of pgpool on different nodes (but not in
watchdogmode). In this case it is not possible to give the responsablity to pgpool to do the failover otherwise both
instanceswould try to do the failover (they don't know about each other, no watchdog mode). So I have to set
failover_commandparameter to an empty string and let repmgr do the automatic failover. I have tested and I believe it
isok: in case of master failure repmgr does the promote, pgpool will not do the failover but will constantly try to
reconnectto a new master and as soon as repmgr has done the promotion it is OK. 
> The important think is that when pgpool starts I have to read the state of the cluster in the repmgr metadata table
(repl_nodes)and generate the /tmp/pgpool_status file based on that. (I even have to do a promote_node in the edge case
thatthe master is down and was brought down when pgpool also was down).   
> But the inconvenient is that if pgpool fails, let's say because the docker node of pgpool is stopped, then docker
swarmrestart the service on another node but this can take a few seconds. So I would prefer to have multiple pgpool. I
didnot have the time right now to test such a scenario but I am interested to know if it is possible. 
> Pierre
>
>     On Wednesday, November 15, 2017, 11:02:31 AM GMT+1, Vikas Sharma <shavikas@gmail.com> wrote:
>
>  Thank you Tatsuo for the Reply, I will post this in list you mentioned.
>
> By Master-Master, I meant two pgpool servers both active accepting connections from Application, If one Pgpool Server
becomesunlavailable other still accepting connections. At an Ideal time, both Pgpool instances on each PgPool server
workingactively.  The database in this case will be 1 master and 2 slaves, replicated by streaming replication. 
>
> Regards
> VS
>
> On 14 November 2017 at 22:56, Tatsuo Ishii <ishii@sraoss.co.jp> wrote:
>
>> Hello There,
>>
>> I need to setup two PGPool Servers in Master-Master mode, First thing I
>> want to know, is it possible?
>>
>> I know we can setup 2 PGPool servers in master-slave mode using watchdog.
>> http://www.pgpool.net/pgpool- web/contrib_docs/watchdog_ master_slave_3.3/en.html
>>
>> Could anyone please enlighten me and any workaround for this?
>>
>> Regards
>> VS
>
>> I need to setup two PGPool Servers in Master-Master mode, First thing I
>> want to know, is it possible?
>>
>> I know we can setup 2 PGPool servers in master-slave mode using watchdog.
>> http://www.pgpool.net/pgpool- web/contrib_docs/watchdog_ master_slave_3.3/en.html
>>
>> Could anyone please enlighten me and any workaround for this?
>
> This is not an appropriate list. You should go to the Pgpool mailing
> list:
>
> https://www.pgpool.net/ mailman/listinfo/pgpool- general
>
> Anyway... I am not sure what you mean by "Master-Master mode" but if
> that means you want to issue write queries by connecting to any of
> Pgpool-II port, it's already possible in the set up you have created.
>
> Best regards,
> --
> Tatsuo Ishii
> SRA OSS, Inc. Japan
> English: http://www.sraoss.co.jp/index_ en.php
> Japanese:http://www.sraoss.co. jp
>
>
>


--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general

pgsql-general by date:

Previous
From: Matt Zagrabelny
Date:
Subject: [GENERAL] expanded auto and header linestyle double
Next
From: Rory Falloon
Date:
Subject: [GENERAL] Setting up replication slave on remote high latency host