Quantum 3.1.3 User Manual

Chapter 6  Managing the File System

Working With Stripe Groups

StorNext User’s Guide

112

A stripe group contains definitions about read and write permissions, 

real time I/O constraints, a stripe breadth definition, multi-pathing 

methodology, and an affinity association. A file system can contain 

multiple stripe groups. 
Stripe groups are bound together to create a StorNext file system. When 

data is written into the file system, two critical things happen. 
First, data is separated from metadata. Metadata operations are typically 

small and random, and they require a lot of head movement on disks. 

Data however, tends to be written in large sequential patterns with less 

head movement. By separating data and metadata, thrashing is 

minimized and performance is maximized. 
The second critical thing that can happen when data is written to the file 

system is file steering. Stripe groups can be different sizes and categories 

of disk, so you could have a file system with one stripe group of SATA 

disk intended for proxy files or temporary storage (i.e., less critical data 

on lower duty cycle, slower disk). You might have another larger fibre 

channel stripe group for storing raw content (i.e., high value data, on 

higher duty cycle, higher performance disk). 
Getting data to specific stripe groups is accomplished using affinities, a 

mapping that ties a directory in the file system to a specific stripe group. 

When you write files to a directory, StorNext uses affinities to 

transparently write those files to the desired disk type. This means you 

don't have to have two or more separate file systems (e.g., one for critical 

data and another for temporary data). Instead, you have a single 

namespace for simplified management.
Working with a stripe group consists of these tasks:

•

•

•

6

Use this procedure to add a stripe group.

From the SNFS home page, choose Stripe Groups from the Config 

menu. The Configure Stripe Groups

 

screen appears.