Обсуждение: Problem with replication

Поиск
Список
Период
Сортировка

Problem with replication

От
Douglas Reed
Дата:
Hi

We are running Pg Version 9.6

On OS Centos 7 running on a VM (Nutanix)

We are trying to build a slave server on this platform. The relevant IP entry for replication has been placed in pg_hba.conf on the master. We run pg_basebackup successfully and then set up recovery.conf etc and re-start 
the slave database. It seems to process the WALs and then enter the read only state when it then generates the following errors;

Mar 23 15:22:48 xds-data1-temp postgres[13072]: [2-1] user=,db=,app=client= LOG:  database system is ready to accept read only connections
Mar 23 15:22:49 xds-data1-temp postgres[13083]: [3-1] user=[unknown],db=[unknown],app=[unknown]client=[local] LOG:  connection received: host=[local]
Mar 23 15:22:49 xds-data1-temp postgres[13083]: [4-1] user=postgres,db=[unknown],app=[unknown]client=[local] FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off
Mar 23 15:22:49 xds-data1-temp postgres[13082]: [3-1] user=,db=,app=client= FATAL:  could not connect to the primary server: FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off
Mar 23 15:22:49 xds-data1-temp postgres[13082]: [3-2]
Mar 23 15:22:49 xds-data1-temp postgres[13085]: [3-1] user=[unknown],db=[unknown],app=[unknown]client=[local] LOG:  connection received: host=[local]
Mar 23 15:22:49 xds-data1-temp postgres[13085]: [4-1] user=postgres,db=[unknown],app=[unknown]client=[local] FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off
Mar 23 15:22:49 xds-data1-temp postgres[13084]: [3-1] user=,db=,app=client= FATAL:  could not connect to the primary server: FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off

We know that the entry is in the hba and have checked for typos, but one thing that does stick out with one line on the message is; 

    FATAL:  no pg_hba.conf entry for replication connection from host "[local]" 

Other messages I have seen in the past would normally have an IP address where is says "[local]"

Any ideas as to where we might be going wrong?    


Many thanks





Doug Reed 
dougreed765@yahoo.com 
07973-132664
https://uk.linkedin.com/pub/douglas-reed/33/326/2b

Re: Problem with replication

От
Holger Jakobs
Дата:

Hi Douglas,

the connection attempt has obviously been started via a "local" connection, using a Unix socket instead of TCP/IP. Therefore, no IP address is involved.

Is the slave db on the same machine?

Regards,

Holger

Am 23.03.20 um 19:30 schrieb Douglas Reed:
Hi

We are running Pg Version 9.6

On OS Centos 7 running on a VM (Nutanix)

We are trying to build a slave server on this platform. The relevant IP entry for replication has been placed in pg_hba.conf on the master. We run pg_basebackup successfully and then set up recovery.conf etc and re-start 
the slave database. It seems to process the WALs and then enter the read only state when it then generates the following errors;

Mar 23 15:22:48 xds-data1-temp postgres[13072]: [2-1] user=,db=,app=client= LOG:  database system is ready to accept read only connections
Mar 23 15:22:49 xds-data1-temp postgres[13083]: [3-1] user=[unknown],db=[unknown],app=[unknown]client=[local] LOG:  connection received: host=[local]
Mar 23 15:22:49 xds-data1-temp postgres[13083]: [4-1] user=postgres,db=[unknown],app=[unknown]client=[local] FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off
Mar 23 15:22:49 xds-data1-temp postgres[13082]: [3-1] user=,db=,app=client= FATAL:  could not connect to the primary server: FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off
Mar 23 15:22:49 xds-data1-temp postgres[13082]: [3-2]
Mar 23 15:22:49 xds-data1-temp postgres[13085]: [3-1] user=[unknown],db=[unknown],app=[unknown]client=[local] LOG:  connection received: host=[local]
Mar 23 15:22:49 xds-data1-temp postgres[13085]: [4-1] user=postgres,db=[unknown],app=[unknown]client=[local] FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off
Mar 23 15:22:49 xds-data1-temp postgres[13084]: [3-1] user=,db=,app=client= FATAL:  could not connect to the primary server: FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off

We know that the entry is in the hba and have checked for typos, but one thing that does stick out with one line on the message is; 

    FATAL:  no pg_hba.conf entry for replication connection from host "[local]" 

Other messages I have seen in the past would normally have an IP address where is says "[local]"

Any ideas as to where we might be going wrong?    


Many thanks





-- 
Holger Jakobs, Bergisch Gladbach, Tel. +49-178-9759012

AW: Problem with replication

От
"Tan, Mein Sin"
Дата:

Hello Douglas,

 

Take a look at the pg_hba.conf on the Replication DB.

 

Best regards,

Mein Sin

 

 

Von: Douglas Reed <dougreed765@yahoo.com>
Gesendet: Montag, 23. März 2020 19:31
An: pgsql-admin@postgresql.org
Cc: Matthew Couper <mcouper@fsbtech.com>; Andrew Jobson <andrew@fsbtech.com>
Betreff: Problem with replication

 

Hi

 

We are running Pg Version 9.6

 

On OS Centos 7 running on a VM (Nutanix)

 

We are trying to build a slave server on this platform. The relevant IP entry for replication has been placed in pg_hba.conf on the master. We run pg_basebackup successfully and then set up recovery.conf etc and re-start 

the slave database. It seems to process the WALs and then enter the read only state when it then generates the following errors;

 

Mar 23 15:22:48 xds-data1-temp postgres[13072]: [2-1] user=,db=,app=client= LOG:  database system is ready to accept read only connections

Mar 23 15:22:49 xds-data1-temp postgres[13083]: [3-1] user=[unknown],db=[unknown],app=[unknown]client=[local] LOG:  connection received: host=[local]

Mar 23 15:22:49 xds-data1-temp postgres[13083]: [4-1] user=postgres,db=[unknown],app=[unknown]client=[local] FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off

Mar 23 15:22:49 xds-data1-temp postgres[13082]: [3-1] user=,db=,app=client= FATAL:  could not connect to the primary server: FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off

Mar 23 15:22:49 xds-data1-temp postgres[13082]: [3-2]

Mar 23 15:22:49 xds-data1-temp postgres[13085]: [3-1] user=[unknown],db=[unknown],app=[unknown]client=[local] LOG:  connection received: host=[local]

Mar 23 15:22:49 xds-data1-temp postgres[13085]: [4-1] user=postgres,db=[unknown],app=[unknown]client=[local] FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off

Mar 23 15:22:49 xds-data1-temp postgres[13084]: [3-1] user=,db=,app=client= FATAL:  could not connect to the primary server: FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off

 

We know that the entry is in the hba and have checked for typos, but one thing that does stick out with one line on the message is; 

 

    FATAL:  no pg_hba.conf entry for replication connection from host "[local]" 

 

Other messages I have seen in the past would normally have an IP address where is says "[local]"

 

Any ideas as to where we might be going wrong?    

 

 

Many thanks

 

 

 

 

 

Re: Problem with replication

От
Loai Abdallatif
Дата:
I think the problem related pg_hba.conf file on slave, try to get copy from backup if exist or somewhere else

On Mon, Mar 23, 2020 at 9:20 PM Tan, Mein Sin <MeinSin.Tan@opitz-consulting.com> wrote:

Hello Douglas,

 

Take a look at the pg_hba.conf on the Replication DB.

 

Best regards,

Mein Sin

 

 

Von: Douglas Reed <dougreed765@yahoo.com>
Gesendet: Montag, 23. März 2020 19:31
An: pgsql-admin@postgresql.org
Cc: Matthew Couper <mcouper@fsbtech.com>; Andrew Jobson <andrew@fsbtech.com>
Betreff: Problem with replication

 

Hi

 

We are running Pg Version 9.6

 

On OS Centos 7 running on a VM (Nutanix)

 

We are trying to build a slave server on this platform. The relevant IP entry for replication has been placed in pg_hba.conf on the master. We run pg_basebackup successfully and then set up recovery.conf etc and re-start 

the slave database. It seems to process the WALs and then enter the read only state when it then generates the following errors;

 

Mar 23 15:22:48 xds-data1-temp postgres[13072]: [2-1] user=,db=,app=client= LOG:  database system is ready to accept read only connections

Mar 23 15:22:49 xds-data1-temp postgres[13083]: [3-1] user=[unknown],db=[unknown],app=[unknown]client=[local] LOG:  connection received: host=[local]

Mar 23 15:22:49 xds-data1-temp postgres[13083]: [4-1] user=postgres,db=[unknown],app=[unknown]client=[local] FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off

Mar 23 15:22:49 xds-data1-temp postgres[13082]: [3-1] user=,db=,app=client= FATAL:  could not connect to the primary server: FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off

Mar 23 15:22:49 xds-data1-temp postgres[13082]: [3-2]

Mar 23 15:22:49 xds-data1-temp postgres[13085]: [3-1] user=[unknown],db=[unknown],app=[unknown]client=[local] LOG:  connection received: host=[local]

Mar 23 15:22:49 xds-data1-temp postgres[13085]: [4-1] user=postgres,db=[unknown],app=[unknown]client=[local] FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off

Mar 23 15:22:49 xds-data1-temp postgres[13084]: [3-1] user=,db=,app=client= FATAL:  could not connect to the primary server: FATAL:  no pg_hba.conf entry for replication connection from host "[local]", user "postgres", SSL off

 

We know that the entry is in the hba and have checked for typos, but one thing that does stick out with one line on the message is; 

 

    FATAL:  no pg_hba.conf entry for replication connection from host "[local]" 

 

Other messages I have seen in the past would normally have an IP address where is says "[local]"

 

Any ideas as to where we might be going wrong?    

 

 

Many thanks

 

 

 

 

 



--
Best Regards,