Re: [HACKERS] Odd listen_addresses behavior - Mailing list pgsql-hackers

From Tom Lane
Subject Re: [HACKERS] Odd listen_addresses behavior
Date
Msg-id 4153.1489607873@sss.pgh.pa.us
Whole thread Raw
In response to [HACKERS] Odd listen_addresses behavior  ("Joshua D. Drake" <jd@commandprompt.com>)
Responses Re: [HACKERS] Odd listen_addresses behavior  ("Joshua D. Drake" <jd@commandprompt.com>)
List pgsql-hackers
"Joshua D. Drake" <jd@commandprompt.com> writes:
> jd@jd-wks:~/snap/postgresql96/common$ grep listen_addresses
> data/postgresql.conf
> listen_addresses = '192*'        # what IP address(es) to listen on;

> -- I wasn't actually expecting the above to work. I was just testing.

Fails as expected for me:

$ postgres --listen-addresses='192*'
2017-03-15 15:50:11.024 EDT [3852] LOG:  could not translate host name "192*", service "5432" to address: Name or
servicenot known 
2017-03-15 15:50:11.024 EDT [3852] WARNING:  could not create listen socket for "192*"
2017-03-15 15:50:11.024 EDT [3852] FATAL:  could not create any TCP/IP sockets
2017-03-15 15:50:11.024 EDT [3852] LOG:  database system is shut down


> postgres=# show listen_addresses ;
>   listen_addresses
> ------------------
>   *
> (1 row)

I'm suspicious that you have an override of listen_addresses somewhere ---
for instance, the "-i" postmaster command line switch effectively is
--listen-addresses='*'.  Even if you had a version of getnameinfo() that
failed to complain about '192*', that would not cause the recorded value
of the string GUC to silently transmogrify into something else.  You might
look into pg_settings to see where it says that value of listen_addresses
came from.
        regards, tom lane



pgsql-hackers by date:

Previous
From: Ashutosh Sharma
Date:
Subject: Re: [HACKERS] Microvacuum support for Hash Index
Next
From: Andres Freund
Date:
Subject: Re: [HACKERS] WIP: Faster Expression Processing v4