Currently the Product Library H2 database is used in embedded mode, which allows only 1 connection.
The error is thrown when a new connection is required (after an existing one).
However, this should not impact the gpt (it appears because it’s part of SNAP initialization).
Server mode should be used for H2 database, for allowing multiple connections.
I will create a JIRA ticket for this.
(https://senbox.atlassian.net/browse/SNAP-1359)