[pgpool-general: 4118] Re: pgpool show both nodes as standby after a failover

[Christophe Le Roux]

Hi,

First of all, thank you for your help.
My search_primary_node_timeout parameter was too low, this was the first issue.

The second issue was a little strange, because all is working fine since I deleted the pgpool_status file ( with pgpool service stopped ) on both nodes.
It seem that on one node the file was not updated and gave a bad information to the 'show pool_nodes' pgpool command.

Did you ever see that ?


Regards,

LE ROUX Christophe

-----Message d'origine-----
De : Tatsuo Ishii [mailto:ishii at postgresql.org] 
Envoyé : jeudi 15 octobre 2015 11:01
À : Christophe Le Roux <christophe.le.roux at fr.clara.net>
Cc : pgpool-general at pgpool.net
Objet : Re: [pgpool-general: 4116] Re: pgpool show both nodes as standby after a failover

I suggest you to diagnose the problem is related to watchdog or not first. I mean you disable watchdog and try the fail over to see if things are getting better.

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

> I can provide more information if you need them.
> 
> Can someone help me ?
> 
> 
> Regards,
> 
> LE ROUX Christophe
> 
> De : pgpool-general-bounces at pgpool.net 
> [mailto:pgpool-general-bounces at pgpool.net] De la part de Christophe Le 
> Roux Envoyé : mardi 13 octobre 2015 11:42 À : 
> pgpool-general at pgpool.net Objet : [pgpool-general: 4113] pgpool show 
> both nodes as standby after a failover
> 
> Hello,
> 
> I'm trying to setup a dual node HA architecture with postgresql 9.3 (with streaming replication) and pgpool 3.4.3 (with watchdog).
> 
> Before the failover, pgpool show the right configuration :
> 
> psql -hserv-e-pgpool-vip.mydomain.net -p 9999 -U postgres -c 'show pool_nodes'
> node_id |  hostname   | port | status | lb_weight |  role
> ---------+-------------+------+--------+-----------+---------
> 0       | serv-ed01 | 5432 | 2      | 0.500000  | primary
> 1       | serv-ed02 | 5432 | 2      | 0.500000  | standby
> (2 rows)
> 
> 
> After the failover, show pool_nodes display that :
> 
> psql -hserv-e-pgpool-vip.mydomain.net -p 9999 -U postgres -c 'show pool_nodes'
> node_id |  hostname   | port | status | lb_weight |  role
> ---------+-------------+------+--------+-----------+---------
> 0       | serv-ed01 | 5432 | 3      | 0.500000  | standby
> 1       | serv-ed02 | 5432 | 2      | 0.500000  | standby
> (2 rows)
> 
> 
> I have configured sr_check* like this :
> 
> # - Streaming -
> 
> sr_check_period = 10
>                                    # Streaming replication check period
>                                    # Disabled (0) by default 
> sr_check_user = 'postgres'
>                                    # Streaming replication check user
>                                    # This is neccessary even if you disable streaming
>                                    # replication delay check by 
> sr_check_period = 0 sr_check_password = 'xxx'
>                                    # Password for streaming 
> replication check user delay_threshold = 10000000
>                                    # Threshold before not dispatching query to standby node
>                                    # Unit is in bytes
>                                    # Disabled (0) by default
> 
> 
> I don't know where is my mistake, can you help me please ?
> 
> 
> Ps: the failover script is triggered and do his job normaly (so the serv-ed02 is really a new primary).
> 
> 
> Regards,
> 
> 
> LE ROUX Christophe