Patrick De Visschere wrote:
> Dave,
>
> The table contains 3 floats, 1 timestamp and 1 integer as the PK. (I've
> defined the troublesome database and table starting with pgadmin 1.6.2
> and since I experienced a lot of crashes I've switched to the snapshots).
> I can access the same database and table with navicat Postgresql and
> with webobjects without problems. Therefore I think that there is
> nothing wrong with the database itself.
>
> I have defined a 2nd table with the same fields and the same data and
> there are no problems with that one.
> So, now I have 2 tables which (to me) look exactly the same, except for
> the order of the fields: one crashes when editing the other not.
>
> [pgAdmin also crashes when making a backup (also if only the good table
> is selected).]
>
> In the pgadmin.log file the last entry (before the crash) is the version
> query:
>
> 2007-02-19 14:16:35 QUERY : Scalar query (127.0.0.1:5432): SELECT
> version();
> 2007-02-19 14:16:35 QUERY : Query result: PostgreSQL 8.2.3 on
> i386-apple-darwin8.9.1, compiled by GCC i686-apple-darwin8-gcc-4.0.1
> (GCC) 4.0.1 (Apple Computer, Inc. build 5367)
>
> normally then the update should follow.
>
> Is there anything else I can do to pinpoint the problem?
I've heard anecdotal reports of crashes (not just on OSX) like this
before - noone has been able to provide a decent test case, and form
what I recall, the problem has always vanished if the table has been
recreated.
Is it possible to get a filesystem level copy of your database to play
with? Perhaps you could copy the data directory (when PostgreSQL is
stopped), and start a postmaster on the copy, then remove everything
except the table that triggers the bug, stop the postmaster again and
send me the entire data directory.
Thanks, Dave.
