[pgpool-general: 6805] Re: Slave Down from pgpool

[Tatsuo Ishii]

First of all, 3.3.4 is not maintained any more. Please use Pgpool-II
versions maintained by community. (see
https://pgpool.net/mediawiki/index.php/EOL_information more info on
EOL).

I see you use repmgr which in my understanding it does failover, while
Pgpool-II does failover as well. So I suspect there might be conflict
between them in terms of failover.

>From what I understand, Pgpool-II surely detects connection problem
with standby. Maybe repmgr does something to standby which makes
Pgpool-II thinks that there's a problem with standby?

> Thanks Tatsuo
> 
> here is the reply:
> 1- pgpool-II version 3.3.4 (tokakiboshi)
> 2-  psql -U pgpool --dbname=postgres --host 10.10.0.2 -c "show pool_nodes"
>  node_id |  hostname  | port | status | lb_weight |  role
> ---------+------------+------+--------+-----------+---------
>  0       | pgserver01 | 3432 | 2      | 0.500000  | primary
>  1       | pgserver02 | 3432 | 3      | 0.500000  | standby
> (2 rows)
> 3- pgpool.conf - attached
> 
> Kind Regards,
> 
> 
> On Tue, Dec 31, 2019 at 3:18 AM Tatsuo Ishii <ishii at sraoss.co.jp> wrote:
> 
>> Looks strange. It seems every time client_idle_limit is hit, failover
>> happens, or vice versa. I need more info. Can you share:
>>
>> - exact Pgpool-II version
>> - output from "show pool_nodes"
>> - pgpool.conf
>>
>> Best regards,
>> --
>> Tatsuo Ishii
>> SRA OSS, Inc. Japan
>> English: http://www.sraoss.co.jp/index_en.php
>> Japanese:http://www.sraoss.co.jp
>>
>> > Dear all
>> > i have setup two nodes of postgres 9.4 with streaming replication and
>> they
>> > work fine as below command:
>> >
>> >  $repmgr -f repmgr/repmgr.conf cluster show
>> > Role      | Connection String
>> > * master  | host=10.10.0.2 (pgserver01)  port=3432 user=repmgr
>> dbname=repmgr
>> >   standby | host=10.10.0.3(pgserver02)  port=3432 user=repmgr
>> dbname=repmgr
>> >
>> > and sometimes the slave being deatatched from pgpool while its running
>> and
>> > accepting data from master node but from pgadmin its beeing deattached
>> and
>> > the error below appears in pgpool logs :
>> >
>> > any one have an idea
>> >
>> >
>> > ERROR: pid 1770: health check failed. 1 th host pgserver02 at port 3432
>> is
>> > down
>> > ec 30 12:57:26 dataserver01 pgpool[1770]: 2019-12-30 12:57:26 LOG:   pid
>> > 40056: pool_process_query: child connection forced to terminate due to
>> > client_idle_limit (180) reached
>> > Dec 30 12:57:27 dataserver01 pgpool[1770]: 2019-12-30 12:57:27 LOG:   pid
>> > 9877: connect_inet_domain_socket: select() timed out. retrying...
>> > Dec 30 12:57:27 dataserver01 pgpool[1770]: 2019-12-30 12:57:27 LOG:   pid
>> > 1770: connect_inet_domain_socket: select() interrupted
>> > Dec 30 12:57:27 dataserver01 pgpool[1770]: 2019-12-30 12:57:27 ERROR: pid
>> > 1770: make_persistent_db_connection: connection to pgserver02(3432)
>> failed
>> > Dec 30 12:57:27 dataserver01 pgpool[1770]: 2019-12-30 12:57:27 ERROR: pid
>> > 1770: health check failed. 1 th host pgserver02 at port 3432 is down
>> > Dec 30 12:57:27 dataserver01 pgpool[1770]: 2019-12-30 12:57:27 LOG:   pid
>> > 40604: connect_inet_domain_socket: select() timed out. retrying...
>> > Dec 30 12:57:28 dataserver01 pgpool[1770]: 2019-12-30 12:57:28 LOG:   pid
>> > 40743: pool_process_query: child connection forced to terminate due to
>> > client_idle_limit (180) reached
>> > Dec 30 12:57:28 dataserver01 pgpool[1770]: 2019-12-30 12:57:28 LOG:   pid
>> > 43528: pool_process_query: child connection forced to terminate due to
>> > client_idle_limit (180) reached
>> > Dec 30 12:57:29 dataserver01 pgpool[1770]: 2019-12-30 12:57:29 LOG:   pid
>> > 42497: pool_process_query: child connection forced to terminate due to
>> > client_idle_limit (180) reached
>> > Dec 30 12:57:30 dataserver01 pgpool[1770]: 2019-12-30 12:57:30 LOG:   pid
>> > 42306: pool_process_query: child connection forced to terminate due to
>> > client_idle_limit (180) reached
>> > Dec 30 12:57:37 dataserver01 pgpool[1770]: 2019-12-30 12:57:37 LOG:   pid
>> > 9877: connect_inet_domain_socket: select() timed out. retrying...
>> > Dec 30 12:57:37 dataserver01 pgpool[1770]: 2019-12-30 12:57:37 LOG:   pid
>> > 40604: connect_inet_domain_socket: select() timed out. retrying...
>> > Dec 30 12:57:38 dataserver01 pgpool[1770]: 2019-12-30 12:57:38 LOG:   pid
>> > 44622: pool_process_query: child connection forced to terminate due to
>> > client_idle_limit (180) reached
>> > Dec 30 12:57:39 dataserver01 pgpool[1770]: 2019-12-30 12:57:39 LOG:   pid
>> > 42245: pool_process_query: child connection forced to terminate due to
>> > client_idle_limi
>>