All right, Brian, I followed your recommendations, by tweaking the parameters of the db. As you anticipated, the shmmax param got exceeded. I changed that param to 1GB in the /etc/sysctl.conf file - hopefully it will be loaded properly on next reboot, do you know anything about that (I'm running Ubuntu 10 or 11)?
And then I managed to restart the db.

Tomorrow, when my colleague is back, we will test out the new performance and let you know, how it is.