[Zope-DB] ZPsycopgDA Connections and Refreshed Zope Products

Federico Di Gregorio fog@initd.org
23 Mar 2003 17:59:10 +0100


--=-dUKL+yyzTuDV9U+sjeJS
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

Il sab, 2003-03-22 alle 03:58, Federico Sevilla III ha scritto:

> product is using, recently even hitting ~70 connections. I could not
> explain this, but through some searching in the archives read a
> message[1] to you and the Zope-DB list by Andrew Veitch, describing a
> problem that I have found to be exactly like mine.
>=20
> [1] http://marc.free.net.ph/message/BA5D8B8F.86BA%andrew@logicalprogressi=
on.net.html
>=20
> To verify, I restarted Zope, effectively closing all connections to
> PostgreSQL. Then I accessed the instance of my product, initating a
> connection. I did a number of other stuff, and still there was exactly
> one connection. Then I updated the timestamp of a file in my product's
> source directory, recompiled the .pyc and .pyo files, and refreshed my
> product. The next access to an instance of my product created a new
> connection to the PostgreSQL database.

this is the very first time someone reports a way to reproduce the
problem. note that i am not even sure it is really a problem because
zope's object cache and threads can make the number of connections very
high. can you send me a cut-down version of your product and a checklist
to expose the problem. the minimum to start debugging.

thank you very much,
federico



--=-dUKL+yyzTuDV9U+sjeJS
Content-Type: application/pgp-signature; name=signature.asc

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQA+fefdvcCgrgZGjesRAgcIAKCnPRl50Turgh3VVNPAt+Vb0mOxjQCgwLSL
HMx2rQRzVsV5wK4lFHYRQ9E=
=JhAg
-----END PGP SIGNATURE-----

--=-dUKL+yyzTuDV9U+sjeJS--