Tom Lane wrote:
Michael Akinde <michael.akinde@met.no> writes:
Tom Lane wrote:
In the past we've seen this type of thing caused by multithreaded
client programs in which more than one thread tried to use the same
PGconn object without adequate interlocking.
Our application is single-threaded, so it seems unlikely that we are
running into a problem with that.
Well, maybe you've found an actual bug then; but without a test case
that other people can poke at, it's hard to investigate.
The only thing I can think of right now, is that we are running a
Postgres 8.3 on Debian Etch (so a backported debian package), whereas
the libraries linked into our application are older library version
(libpq4 and libpqxx 2.6.8).
Older libpq versions should work fine with an 8.3 server --- if they
don't, that's a bug in itself. I do not know the status of libpqxx
though; you might want to check whether there are known bugs in that
version.
There are a few known bugs in pqxx that *might* relate to our code; for instance, we use prepared statements, and this is an area that has seen a good deal of work from 2.6.8 to 2.6.9.
The test work at the moment is hampered by our server section needing to physically move some of our development & test servers this week, but if upgrading pqxx doesn't seem to eliminate the problem, I'll do some work to see if I can recreate the problem in an isolated setting.
Regards,
Michael A.