SciDB

Introduction

SciDB (http://www.scidb.org/) is a parallel database system we are evaluating as part of the Parallel Database Testbed project. 

Requesting Access

To request access to NERSC SciDB please email consult@nersc.gov, briefly explain your use case, expected data volume, sample operations, and current solution. 

Using SciDB

login to carver:

ssh carver.nersc.gov

then login to c0130-ib

ssh c0130-ib.nersc.gov

use iquery:

iquery
AQL% select * from list('arrays'); 

SciDB Manual

Please refer to the SciDB User Manual for usage of SciDB (Attached). 

Storage Space

There is Global HOME/SCRATCH on the SciDB nodes. However the Global PROJECT is mounted. You can load data directly from your project directory. Fo temporary storage space, you can create a subdirectory in /global/project/projectdirs/paralleldb/users. 

You can transfer your data from global file systems using rsync. 

Sharing the System With Other Users

SciDB currently doesn't provide user level access control, this means you can accidentally delete other user's data. 

Naming Your Arrays

To distinguish different users, please name your array with your user name as a prefix. For example, if your NIM username is testuser, and the desired array name is myarray, please name your array "testuser_myarray"

Important Notice

This is a testbed system, please backup your data and queries. 

Questions

Please email consult@nersc.gov