I had this posted in another thread before, where the problem seems to be
related, but it was apparently not. so here's my problem again:

i ran into the following issue
with jaybird and glassfish connection pooling just yesterday. atm im trying to
setup a testscenario.
 the following happened:
we tried to deploy a webapplication (db access via spring/hibernate) that is
normally deployed on tomcat to glassfish 2.1. The first thing we noticed was
the infamous "Could not obtain connection during blocking timeout (5000 ms)"
exception when our application started to use more than 10 connections at a
time. tomcat running on the same machine with the same jaybird against the
same database did allocate its connections just fine.
Firts thing we checked were the poolsettings ofc. Nothing special there. When
i raised the minimum number of connections in pool above 10, the creation of
the pool itself would fail, otherwise the error would occur when our
application requested more than 10 connections from pool.
Next suspect was xinetd, since we were running classic server.i changed the
settings in et/xinetd.d/firebird, then replaced classic with superserver.
but the problem remained.
at the moment i am trying to setup a testcase on my developer machine and plan
to debug in to glassfishs and jaybirds sources to see whats going on.

lutz