Re: what happens when you issue ALTER SERVER in a hot environment?

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: "David G(dot) Johnston" <david(dot)g(dot)johnston(at)gmail(dot)com>
Cc: AC Gomez <antklc(at)gmail(dot)com>, "pgsql-generallists(dot)postgresql(dot)org" <pgsql-general(at)lists(dot)postgresql(dot)org>
Subject: Re: what happens when you issue ALTER SERVER in a hot environment?
Date: 2020-04-07 04:53:11
Message-ID: 25511.1586235191@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

"David G. Johnston" <david(dot)g(dot)johnston(at)gmail(dot)com> writes:
> On Monday, April 6, 2020, AC Gomez <antklc(at)gmail(dot)com> wrote:
>> If you issue an ALTER SERVER command and there are active connections
>> with that server in use or new ones are coming in, what happens? Docs on
>> this command say nothing regarding active processing using the server
>> context and changes to it. So I assume it's just handled.

> Not sure if there are exceptions but assume that nothing external will
> change your current active session’s settings out from underneath you.

Couple of comments here:

* ALTER SERVER changes nothing until "pg_ctl reload" or similar is
issued. The same goes for manually editing the config file (which
is more or less the same thing, though on a different file).

* Once you do issue a reload, individual sessions will adopt the
new setting reasonably promptly, though not necessarily all at the
same instant (assuming it's a setting that doesn't require a
postmaster restart to take effect, in which case nothing happens).

* A session that has a higher-priority source for the setting of
a given GUC, such as a locally-issued SET, is not going to adopt
the new setting from the config file. But it *will* adopt that
setting on-the-fly if it has no other source of the setting.
The priority rules are explained in the fine manual, IIRC.
Possibly some clarification is needed?

>> For example if you alter user/password, I assume that as long as prior
>> user password is still good that actively running processes will keep going.

> If you alter a password the old one is by definition no longer good...but
> it doesn’t matter because authentication only happens once - during login.

Right. Also, passwords are *not* managed through the GUC (server
parameter) mechanism; what I just said about GUCs doesn't apply
to them.

regards, tom lane

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Pavel Stehule 2020-04-07 05:03:57 Re: How to unnest nested arrays
Previous Message David G. Johnston 2020-04-07 04:29:37 Re: what happens when you issue ALTER SERVER in a hot environment?