Re: Cache invalidation after authentication (on-the-fly role creation) - Mailing list pgsql-hackers

From Thomas Munro
Subject Re: Cache invalidation after authentication (on-the-fly role creation)
Date
Msg-id CAEepm=0jV_o4_ZBZ=F0OT3BLk7O7q7Er-7gFfnGq5c-ERpH6Uw@mail.gmail.com
Whole thread Raw
In response to Re: Cache invalidation after authentication (on-the-fly rolecreation)  (Andres Freund <andres@anarazel.de>)
List pgsql-hackers
On Wed, Jul 4, 2018 at 4:35 PM, Michael Paquier <michael@paquier.xyz> wrote:
> I wanted to comment on that this morning but forgot as my mind was
> driven away by another problem.  What if you used the Julien-Rouhaud's
> method of a custom script with only ";" used as query and -c?  This
> won't run any queries, and will stress authentication.

Ok, I tried "pgbench -c 8 -j 8 -T 60 --connect -f empty.sql postgres"
where empty.sql contains just ";", and the numbers were noisy but
around 1160 TPS with or without a patch.

On Wed, Jul 4, 2018 at 4:35 PM, Andres Freund <andres@anarazel.de> wrote:
> On 2018-07-04 16:25:18 +1200, Thomas Munro wrote:
>>               PerformAuthentication(MyProcPort);
>> +             AcceptInvalidationMessages();
>>               InitializeSessionUserId(username, useroid);
>
> FWIW, a comment explaining why it's done there seems appropriate.

Done.

-- 
Thomas Munro
http://www.enterprisedb.com

Attachment

pgsql-hackers by date:

Previous
From: Peter Eisentraut
Date:
Subject: Re: CREATE TABLE .. LIKE .. EXCLUDING documentation
Next
From: Pavel Raiskup
Date:
Subject: [PATCH] btree_gist: fix union implementation for variable length columns