Обсуждение: [GENERAL] Weird periodical pg log

Every single day at exactly the same hour I get this in my pg_log:


< 2017-05-15 17:00:29.517 CEST >FATAL:  pg_hba.conf rejects connection
for host "...", user "postgres", database "template1", SSL off
< 2017-05-15 17:00:29.571 CEST >WARNING:  archive_mode enabled, yet
archive_command is not set
< 2017-05-15 17:00:41.859 CEST >LOG:  received SIGHUP, reloading
configuration files
< 2017-05-15 17:00:41.861 CEST >LOG:  parameter "archive_command"
changed to ""/opt/omni/lbin/pgsqlbar.exe" -stage %p -backup"
< 2017-05-15 17:00:51.865 CEST >LOG:  received SIGHUP, reloading
configuration files
< 2017-05-15 17:00:51.867 CEST >LOG:  parameter "archive_command"
removed from configuration file, reset to default


Postgres 9.5 is running on Centos 7. I checked all cron jobs and
scripts, there is nothing external that would cause this so my guess is
that this is internal Postgres cron job tryin to do "something", I just
don't know what exactly.

We have WAL replication set up, these logs appear on master. Perhaps we
have something off in our conf?


Best regards, cen

On Fri, May 19, 2017 at 10:25:13AM +0200, cen wrote:

> < 2017-05-15 17:00:41.861 CEST >LOG:  parameter "archive_command" changed to ""/opt/omni/lbin/pgsqlbar.exe" -stage %p
-backup"

This is the line that you'll have to base your research on.

Also, you might want to check for a keylogger in the audio driver.

Regards,
Karsten
--
GPG key ID E4071346 @ eu.pool.sks-keyservers.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346

On 05/19/2017 01:25 AM, cen wrote:
> Every single day at exactly the same hour I get this in my pg_log:
>
>
> < 2017-05-15 17:00:29.517 CEST >FATAL:  pg_hba.conf rejects connection
> for host "...", user "postgres", database "template1", SSL off
> < 2017-05-15 17:00:29.571 CEST >WARNING:  archive_mode enabled, yet
> archive_command is not set
> < 2017-05-15 17:00:41.859 CEST >LOG:  received SIGHUP, reloading
> configuration files
> < 2017-05-15 17:00:41.861 CEST >LOG:  parameter "archive_command"
> changed to ""/opt/omni/lbin/pgsqlbar.exe" -stage %p -backup"

Per Karsten's suggestion, the above seems to come from HP Enterprise
software. As a start I would see what pgsqlbar.exe -help says.

> < 2017-05-15 17:00:51.865 CEST >LOG:  received SIGHUP, reloading
> configuration files
> < 2017-05-15 17:00:51.867 CEST >LOG:  parameter "archive_command"
> removed from configuration file, reset to default
>
>
> Postgres 9.5 is running on Centos 7. I checked all cron jobs and
> scripts, there is nothing external that would cause this so my guess is
> that this is internal Postgres cron job tryin to do "something", I just
> don't know what exactly.

It is not an internal Postgres job. Something external to Postgres is
making a change to postgresql.conf reloading the server conf files,
undoing the change and then reloading the server conf files again.

>
> We have WAL replication set up, these logs appear on master. Perhaps we
> have something off in our conf?
>
>
> Best regards, cen
>
>
>


--
Adrian Klaver
adrian.klaver@aklaver.com

On 5/19/2017 1:25 AM, cen wrote:
> < 2017-05-15 17:00:41.861 CEST >LOG:  parameter "archive_command"
> changed to ""/opt/omni/lbin/pgsqlbar.exe" -stage %p -backup"


I believe /opt/omni is the default installation path for HP Data
Protector, formerly known as OmniBack.      That comprehensive backup
system includes database backup capabilities, I'm guessing thats what
you're seeing here.


--
john r pierce, recycling bits in santa cruz

Hmm indeed that seems to be the case. Thanks for the tip!


On 05/19/2017 04:10 PM, John R Pierce wrote:
> On 5/19/2017 1:25 AM, cen wrote:
>> < 2017-05-15 17:00:41.861 CEST >LOG:  parameter "archive_command"
>> changed to ""/opt/omni/lbin/pgsqlbar.exe" -stage %p -backup"
>
>
> I believe /opt/omni is the default installation path for HP Data
> Protector, formerly known as OmniBack.      That comprehensive backup
> system includes database backup capabilities, I'm guessing thats what
> you're seeing here.
>
>