Re: Adding a pg_servername() function - Mailing list pgsql-hackers

From Laetitia Avrot
Subject Re: Adding a pg_servername() function
Date
Msg-id CAB_COdgPR_oXRt=Sf8V+9V74uaJnuT7MmUr4ySfSRa7B5KQUEA@mail.gmail.com
Whole thread Raw
In response to Re: Adding a pg_servername() function  (Christoph Moench-Tegeder <cmt@burggraben.net>)
List pgsql-hackers
Dear Christoph,

Please find my answers below.

Le mer. 9 août 2023 à 22:05, Christoph Moench-Tegeder <cmt@burggraben.net> a écrit :
## GF (phabriz@gmail.com):

And now that I checked it: I do have systems with gethostname()
returning an FQDN, and other systems return the (short) hostname
only.

The return of gethostname() depends on what has been configured. So, yes some people will prefer a FQDN while others will prefer a short hostname. Also, it's a POSIX standard function (see https://pubs.opengroup.org/onlinepubs/9699919799/), so I don't get why getting a FQDN or a short name depending on what people set would be a problem for Postgres while it's not for Linux.
 
And it gets worse when you're talking "container" and
"automatic image deployment". So I believe it's a good thing when
a database does not expose too much of the OS below it...


I respectfully disagree. Containers still have a hostname: for example, by default, docker uses their container id. (see https://docs.docker.com/network/). In fact, the Open Container Initiative defines the hostname as a runtime configuration parameter (see https://github.com/opencontainers/runtime-spec/blob/main/config.md):

Hostname

Example

"hostname": "mrsdalloway"

Have a nice day,

Lætitia

pgsql-hackers by date:

Previous
From: Ashutosh Bapat
Date:
Subject: Re: [BackendXidGetPid] only access allProcs when xid matches
Next
From: Junwang Zhao
Date:
Subject: Re: [BackendXidGetPid] only access allProcs when xid matches