Re: BUG #15759: Lot of idle connection created by "SETextra_float_digits = 3" - Mailing list pgsql-bugs

From anup singh
Subject Re: BUG #15759: Lot of idle connection created by "SETextra_float_digits = 3"
Date
Msg-id CA+i39PzkRwqyuFROeS+a3YQRE2Rg+u6GJ1czRJKnq5RdRmTV8g@mail.gmail.com
Whole thread Raw
In response to Re: BUG #15759: Lot of idle connection created by "SETextra_float_digits = 3"  (Tomas Vondra <tomas.vondra@2ndquadrant.com>)
Responses Re: BUG #15759: Lot of idle connection created by "SETextra_float_digits = 3"  (Francisco Olarte <folarte@peoplecall.com>)
List pgsql-bugs
Thank you very much Tomas for such a nice explanation.
We have Max 100 connection. If you see the screen shot in previous mail, most of the idle connection are taken by "SET extra_float_digits = 3" bit. After sometime application says "fatal error - reached Max connection, 3 connection are reserved for PG".

Regards
Anup


On Tue, Apr 16, 2019, 6:07 PM Tomas Vondra <tomas.vondra@2ndquadrant.com> wrote:
On Tue, Apr 16, 2019 at 05:31:42PM +0530, anup singh wrote:
>Hi Tomas,
>
>Thanks for your response. Kindly see the below screen shot where we have
>lot of idle connection which is reaching max pool size soon, but the actual
>request is only one.
>

I'm pretty sure that's the expected behavior of a connection pool. It
opens a bunch of connections, does some initialization on them (which
likely includes the "SET extra_float_digits = 3" bit) and then keeps them
around so that the application can use them.

You have not really explained how it's killing your application, though.
My guess is it's either not going through the connection pool (and so is
hitting max_connections or something) or it's misconfigured in some other
way. In any case, extra_float_digits is not the root cause here.

regards

--
Tomas Vondra                  http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

pgsql-bugs by date:

Previous
From: PG Bug reporting form
Date:
Subject: BUG #15760: Error to start PGAdmin 4.5
Next
From: Jeff Janes
Date:
Subject: Re: BUG #15758: regr_* return wrong answers for some valid data