D-Cysteine / nesql-server

Server for NESQL

Not Enough SQL Server

The server module for NESQL. The exporter module can be found here. Still WIP.

Instructions

1. Place nesql-server-*.jar in the same directory as the nesql-repository folder that was exported by the exporter. The name must be nesql-repository, at least until a customization option is added.
2. Start the server with java -jar nesql-server-*.jar.
3. Visit http://localhost:8080/index to access the server.
4. You can shut down the server by visiting http://localhost:8080/shutdown, or just with Ctrl-C on the command line.

Configuration

Configurable options are listed in this file. To configure them, make a copy of that file in the same directory as the JAR, modify any options that you like, and delete any options that you don't want to modify.

This is still WIP, and some options (like repository name) don't work yet! The most useful option to configure is probably the SQL query timeout, which defaults to 1 min.

Faster startup

NESQL Server reads from an HSQLDB database database. For speed reasons, NESQL Exporter currently exports these databases in TEXT format. If you'd like to greatly reduce the NESQL Server start-up time, you can convert the exported database to CACHED format, though this will also increase query times. The documentation for this topic can be found here.

To perform the conversion, you will need to open the database with the HSQLDB client, which can be downloaded from the link above (you may want to make a copy of the database first, just to be safe). Then, run this command to list all of the tables in the database:

SELECT TABLE_NAME FROM INFORMATION_SCHEMA.SYSTEM_TABLES WHERE TABLE_TYPE = 'TABLE'

Finally, run this block of commands, with an entry for each table:

SET TABLE <table name> TYPE CACHED
SHUTDOWN COMPACT

For example, to compact the ITEM and FLUID tables, you would run:

SET TABLE ITEM TYPE CACHED
SET TABLE FLUID TYPE CACHED
SHUTDOWN COMPACT

This process is likely to take a while, and may moderately increase disk space usage, but you will see a massive reduction in server start-up time, and a moderate increase in query time. Memory usage will also decrease significantly.

It is also possible to tune the performance of a CACHED database, by configuring how many rows are cached in memory. These parameters can be tuned by executing SQL commands in the HSQLDB client. See hsqldb.cache_rows and hsqldb.cache_size here.