Hi Eric,
of course I also cannot debug your problem
remotely, but here are some ideas of what
effects could possibly cause such strange bugs.
First think about connection-pooling. This will
certainly behave different for the build-in datasource
and that of your container.
So if there would be "uncommitted changes" for
whatever reason in some of the pooled connections,
that would exlain the sort of behaviour you describe.
(if you are using a transactional database like oracle)
This sort of bug is my first guess. So for debugging,
it would be interesting to trace the identity of the
connections actually used.
The other source of strangeness could be if
your conatiners datasource does prepared-statement-caching.
Could you check that?
I'm interested if there are any problems related
to prepared-statement-caching, cause I'm implementing
that for cayennes data source.
regards,
Arndt
Eric Schneider wrote:
> Hi,
>
> Kind of an odd situation I'm having here. I have an application that
> has been implemented with cayenne for 6 months now. Recently, we
> modified the application to use a JNDI datasource. The change was very
> straightforward. Instead of looking up connection information in the
> driver file, the factory attribute in the cayenne.xml file was switched
> to the JNDIDataSourceFactory and the datasource attribute was set to our
> JNDI name, jdbc/mysource.
>
> The behavior I'm seeing is extremely odd, but consistent. Using the
> datasource, certain queries (not all) return no rows. The call doesn't
> throw an exception, just simply returns the standard cayenne debug
> output saying "return 0 rows" as if the table was truncated. A query
> will work perfectly qualified one way, but fails to return anything if
> the qualifiers are slightly different (but also correct).
>
> If I copy the SQL that cayenne generates and paste it into SQL plus, it
> returns records. Also, if I switch back to Cayenne's build in
> connection pool, it works perfectly. For whatever reason, it
> just behaves differently. :-|
>
> Maybe someone that's a little more familiar with the spaghetti innards
> of cayenne would know why this could be happening?
>
> Thanks,
> Eric
-- -- Dr. Arndt Brenschede DIAMOS AG Innovapark Am Limespark 2 65843 SulzbachTel.: +49 (0) 61 96 - 65 06 - 134 Fax: +49 (0) 61 96 - 65 06 - 100 mobile: +49 (0) 151 151 36 134 mailto:arndt.brensched..iamos.com http://www.diamos.com
This archive was generated by hypermail 2.0.0 : Fri Apr 11 2003 - 12:03:15 EDT