Is it possible (now or in the future) to have a separate common postgres instance for the SciDB catalog database? I see the potential within the scidb.py script which uses the $SCIDB_INSTALL_PATH/bin/init-db.sh script with everything able to be specified except for the host. It seems like an extremely easy add to the psql commands within init-db.sh but I am sure something would also have to be done within the config.ini file and respective code that handles the communication internally. Is this possible now or in the future?
In the long term, we plan to remove the dependency on Postgres. So we weren’t planning to make a lot of enhancement to Postgres in the short term.
Do you have an ETA on when the Postgres dependence will be removed? This requirement of the coordinator forces users in to a sort of dedicated hardware situation that will greatly decrease the adoption of SciDB in traditional scheduled HPC environments.