Server mode snap engine

snap
1

Hello,

I am just wondering when you use the graph processing tool (gpt) through the command line, does it by default use the snap engine in server mode? Or does it choose when to do so automatically? I am just wondering if it is possible to set the gpt to run in server mode as an issue I am having seems to be related to server mode of snap engine. i.e. when I run multiple instances of gpt for a command such as TopSarsplit at the same time I get an error which says the solution could be “Possible solutions: close all other connection(s); use the server mode [90020-197]”, but I am unsure of how I use the server mode.

Thanks.

2

The message is related to the product library.
It has been reported here already: database.\r\norg.h2.jdbc.JdbcSQLException: Database may be already in use: “Locked by another process” - snap / Problem Reports - STEP Forum (esa.int)

But to answer your question. The engine is not put into any server mode automatically.

3

From what I remember, GPT doesn’t automatically force SNAP into server mode when you run it from the command line. It will usually start SNAP in desktop mode unless server mode is explicitly enabled. To avoid those connection errors when running multiple instances, you’ll need to launch SNAP with server mode enabled or configure it via the appropriate SNAP parameters before running GPT.

4

Running multiple gpt instances in parallel is usually the dream, but that H2 database lock is such a classic headache. If you don’t actually need the Product Library for your command-line processing, a quick ‘human’ fix is to just disable it in your snap.properties file (set snap.productlibrary.active=false). It usually stops the engine from trying to ping that locked database altogether!