Two Postgres master process are showing - one is on and off

On 8/20/19 1:32 PM, chiru r wrote:

Hi All,

I have observed one of our PostgreSQL DB instance showing two postgres
process on Linux server as highlighted. The second postgres process is
on and off.

Is the on/off on a regular interval?
My guess is a script(cron possibly) is trying to start the server.

We did not find any references in logs.

The information:

sudo /etc/init.d/postgresql11 start
Starting PostgreSQL: ok
aklaver@ranger:~> -2019-08-20 13:40:30.097 PDT-0FATAL: lock file
"postmaster.pid" already exists
-2019-08-20 13:40:30.097 PDT-0HINT: Is another postmaster (PID 1694)
running in data directory "/usr/local/pgsql11/data"?

Is sent to the starting process and does not make it to the server log
as the server is not started(the 'Starting PostgreSQL' notwithstanding).

Please provide your advice and help on this issue.

DB version :  PostgreSQL *9.5.10*

Server version   :  RHEL *7.6*

[postgres@PGSERVER data]$ ps -ef|grep postgres

root       5922 5875  0 Jul31 pts/0    00:00:00 sudo su - postgres

root       5927 5922  0 Jul31 pts/0    00:00:00 su - postgres

postgres   5928   5927  0 Jul31 pts/0    00:00:00 -bash

root      11585  11535  0 11:29 pts/2    00:00:00 sudo su - postgres

root      11606  11585  0 11:29 pts/2    00:00:00 su - postgres

postgres  11607  11606  0 11:29 pts/2    00:00:00 -bash

root      17185  17141  0 Aug19 pts/1    00:00:00 sudo su - postgres

root      17189  17185  0 Aug19 pts/1    00:00:00 su - postgres

postgres  17190  17189  0 Aug19 pts/1    00:00:00 -bash

postgres 33438      1  0 12:41 ?        00:00:03
/u01/postgres/9.5/bin/postgres -D /u02/pgdata01/9.5/data

postgres  33441  33438  0 12:41 ?        00:00:02 postgres: logger process

postgres  33443  33438  0 12:41 ?        00:00:01 postgres: checkpointer
process

postgres  33444  33438  0 12:41 ?        00:00:00 postgres: writer process

postgres  33445  33438  0 12:41 ?        00:00:01 postgres: wal writer
process

postgres  33446  33438  0 12:41 ?        00:00:00 postgres: autovacuum
launcher process

postgres  33447  33438  0 12:41 ?        00:00:01 postgres: stats
collector process

postgres  36660  33438  0 12:47 ?        00:00:00 postgres: sj2 devdb
10.220.177.63(57237) idle

root      39915  39837  0 12:53 pts/3    00:00:00 sudo su - postgres

root      39950  39915  0 12:53 pts/3    00:00:00 su - postgres

postgres  39951  39950  0 12:53 pts/3    00:00:00 -bash

postgres 110181  33438  0 15:30 ? 00:00:00
/u01/postgres/9.5/bin/postgres -D /u02/pgdata01/9.5/data

postgres 110183  11607  0 15:30 pts/2    00:00:00 ps -ef

postgres 110184  11607  0 15:30 pts/2    00:00:00 grep --color=auto postgres

[postgres@SERVER1 data]$

Thanks,

Chiru

--
Adrian Klaver
adrian.klaver@aklaver.com

Thank you Adrian.

Is the on/off on a regular interval?

My guess is a script(cron possibly) is trying to start the server.

Yes, It is regular interval.

No Con-jobs scheduled on server or even remotely.

We did not find any references in logs.

The information:

sudo /etc/init.d/postgresql11 start
Starting PostgreSQL: ok
aklaver@ranger:~> -2019-08-20 13:40:30.097 PDT-0FATAL: lock file
"postmaster.pid" already exists
-2019-08-20 13:40:30.097 PDT-0HINT: Is another postmaster (PID 1694)
running in data directory "/usr/local/pgsql11/data"?

Is sent to the starting process and does not make it to the server log
as the server is not started(the 'Starting PostgreSQL' notwithstanding).

I have verified startup.log and regular postgresql logs also, i did not

found above messages.

Thanks,
Chiru

FWIW, the second process (110181) is a child of the main process (33438). I
can't recall if any maintenance process that gets automatically activated
(autovacuum, checkpoint, stats, etc.) shows with the same name as the main
process.

Cheers,
Steve

On Tue, Aug 20, 2019 at 1:32 PM chiru r <chirupg@gmail.com> wrote:

(sorry for the top-post - blame GMail)

-Steve

På tirsdag 20. august 2019 kl. 22:32:25, skrev chiru r <chirupg@gmail.com
<mailto:chirupg@gmail.com>>:
Hi All,

I have observed one of our PostgreSQL DB instance showing two postgres process
on Linux server as highlighted. The second postgres process is on and off.

We did not find any references in logs.

[...]

postgres 33438 1 0 12:41 ? 00:00:03 /u01/postgres/9.5/bin/postgres -D
/u02/pgdata01/9.5/data

[...]

postgres 110181 33438 0 15:30 ? 00:00:00 /u01/postgres/9.5/bin/postgres -D
/u02/pgdata01/9.5/data
Strange, the second is a child of the first... --
Andreas Joseph Krogh

chiru r wrote:

I have observed one of our PostgreSQL DB instance showing two postgres process on Linux server as highlighted. The second postgres process is on and off.
We did not find any references in logs.

Please provide your advice and help on this issue.

DB version : PostgreSQL 9.5.10
Server version : RHEL 7.6

[postgres@PGSERVER data]$ ps -ef|grep postgres
postgres 33438 1 0 12:41 ? 00:00:03 /u01/postgres/9.5/bin/postgres -D /u02/pgdata01/9.5/data
[...]
postgres 110181 33438 0 15:30 ? 00:00:00 /u01/postgres/9.5/bin/postgres -D /u02/pgdata01/9.5/data

Perhaps it is a new client connection, and the process title just hasn't been changed (yet).

What is your setting for "update_process_title"?

Yours,
Laurenz Albe
--
Cybertec | https://www.cybertec-postgresql.com