Thread: Testing of MVCC

Testing of MVCC

From
Matt Miller
Date:
I want to write some regression tests that confirm the behavior of
multiple connections simultaneously going at the same tables/rows.  Is
there something like this already, e.g. in src/test/regress?  In
particular I want confirm the robustness of some PL/pgSQL functions in a
multi-user environment.  I could probably just bang away from multiple
interactive psql sessions, but I like to script the whole thing, and I'd
like it all to fit neatly into the current "make check" regression
tests.

Re: Testing of MVCC

From
Tom Lane
Date:
Matt Miller <mattm@epx.com> writes:
> I want to write some regression tests that confirm the behavior of
> multiple connections simultaneously going at the same tables/rows.  Is
> there something like this already, e.g. in src/test/regress?

No.  You should consult the pghackers archives --- there have been
discussions in the past about creating a test harness that would support
useful concurrent testing.  No one's gotten around to it yet, but surely
we need one.

            regards, tom lane

Re: Testing of MVCC

From
Karsten Hilbert
Date:
> Matt Miller <mattm@epx.com> writes:
> > I want to write some regression tests that confirm the behavior of
> > multiple connections simultaneously going at the same tables/rows.  Is
> > there something like this already, e.g. in src/test/regress?
>
> No.  You should consult the pghackers archives --- there have been
> discussions in the past about creating a test harness that would support
> useful concurrent testing.  No one's gotten around to it yet, but surely
> we need one.
There's something *somewhat* related here:

 http://savannah.gnu.org/cgi-bin/viewcvs/gnumed/gnumed/gnumed/client/testing/concurrency-torture-test.py

Karsten
--
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346

Re: Testing of MVCC

From
Richard_D_Levine@raytheon.com
Date:
Firebird has MVCC also (they call it multi-generational record architecture
--- MGRA), and may have at least a good test plan, though it may not cover
effects of rules, triggers, functions, and constraints.  Those are the
killer test cases.  I don't have time to look.

http://firebird.sourceforge.net/

Rick

pgsql-general-owner@postgresql.org wrote on 08/09/2005 02:19:56 PM:

> > Matt Miller <mattm@epx.com> writes:
> > > I want to write some regression tests that confirm the behavior of
> > > multiple connections simultaneously going at the same tables/rows.
Is
> > > there something like this already, e.g. in src/test/regress?
> >
> > No.  You should consult the pghackers archives --- there have been
> > discussions in the past about creating a test harness that would
support
> > useful concurrent testing.  No one's gotten around to it yet, but
surely
> > we need one.
> There's something *somewhat* related here:
>
>  http://savannah.gnu.org/cgi-
>
bin/viewcvs/gnumed/gnumed/gnumed/client/testing/concurrency-torture-test.py
>
> Karsten
> --
> GPG key ID E4071346 @ wwwkeys.pgp.net
> E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346
>
> ---------------------------(end of broadcast)---------------------------
> TIP 4: Have you searched our list archives?
>
>                http://archives.postgresql.org