[pgpool-general: 5382] Re: [pgpool] Streaming Replication - Identify Slave Node Status

Lazaro Garcia lazaro3487 at gmail.com
Wed Mar 22 22:29:27 JST 2017

Pgpool triggers failover on this situations:


With health_check functionality. Between intervals it checks the status of backends nodes and makes a failover if a node is detected as down.


fail_over_on_backend_error: If this command is set to on,  Initiates failover when reading/writing to the backend communication socket fails. If set to off, pgpool will report an error and disconnect the session.


Check these configurations.


And start pgpool with –D option for descarting pgpool_status file.




De: Anushka Weerakkodyge [mailto:anushkaw at ceylonit.com] 
Enviado el: miércoles, 22 de marzo de 2017 04:37 a. m.
Para: Lazaro Garcia
CC: pgpool-general at pgpool.net
Asunto: Re: [pgpool-general: 5370] [pgpool] Streaming Replication - Identify Slave Node Status


Hi Lazaro,


I tried failover command in pgpool and was able to stop the replication and make the master server a read only node when the standby is down. But I noticed that this procedure does not work always. Sometimes pgpool does not trigger failover mechanism when the standby node is down and log keeps showing failed to make persistent db connection message. Is there something I'm missing here ?


In the config file I turned on master_slave mode and load_balance mode.








On Fri, Mar 17, 2017 at 5:10 AM, Anushka Weerakkodyge <anushkaw at ceylonit.com <mailto:anushkaw at ceylonit.com> > wrote:

Thank you! I will try this method definitely. By the way, what I really need is to fail the transaction when slave is down. Using this method I can identify the status of the slave node, then will try to make the master node read only until the slave comes online.





On Thu, Mar 16, 2017 at 6:36 PM, Lazaro Garcia <lazaro3487 at gmail.com <mailto:lazaro3487 at gmail.com> > wrote:

When slave goes down, pgpool detects that and triggers a failover, then the master remains active and the slave are marked as down state. On the failover_command of pgpool, you could execute a command that drop the slave from the master synchronous_standby_names and reload the configuration on the master. By this way new transactions can be executed on the master without checking the state of the slave because synchronous replication has been disabled.


If you recover the slave you can attach this server to master again an activate the synchronous replication.





De: pgpool-general-bounces at pgpool.net <mailto:pgpool-general-bounces at pgpool.net>  [mailto:pgpool-general-bounces at pgpool.net <mailto:pgpool-general-bounces at pgpool.net> ] En nombre de Anushka Weerakkodyge
Enviado el: miércoles, 15 de marzo de 2017 05:37 p. m.
Para: pgpool-general at pgpool.net <mailto:pgpool-general at pgpool.net> 
Asunto: [pgpool-general: 5370] [pgpool] Streaming Replication - Identify Slave Node Status


Hi all,


I have a synchronous streaming replication environment which is implemented using inbuilt replication mechanism in postgresql. There is a requirement that I need to fail the data modification transactions on master node if the slave node is down. (ie: transactions should commit when only both servers are up and running) 


Please let me know whether there is a way to identify slave node status and fail the transaction using pgpool ? 






-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20170322/87262bda/attachment-0001.html>

More information about the pgpool-general mailing list