Types of Storage Systems - Netapp

Modular,Integrated

Modular System

A modular system is a filler head attached to one or more disk shelves.
Filer (Fabric Attached Storage) - FAS960
NearStore - R200
Netcache - C6200

Integrated

An Integrated system is one in which the head funtionality is built into the disk shelf.
Filer (Fabric Attached Storage) - FAS270
NetCache - C2100

Netapp Hardware Components / Fundamentals

NVRAM:

NVRAM retains data in the event od power loss.

System Memory:

System Memory is the main memory that holds current programs and data that are in use.

CPU:

The Central Processing unit (CPU) is the man processing chip, or brains of the system.

Power supplies:

Power supplies convert building power into lower voltages required by the system.

System fans:

System fans transfer heat awat from sensitive areas and out of the system.

Expansion adapter cards:

Expansion adapter cards extend the system's control over peripheral devies. An example is the NIC or network interface card.

Ethernet:

The on-board Ethernet connection and network interface card (NIC) provide access to networks and network protocols.

Console:

The console connection is a physical or remote terminal that monitors and controls a Netapp system.

Boot Device :

The system uses the boot device to initialize the system at startup.

Disk Drives:

Two or more disk drives are used to store and retrieve data.

ESH/LRC Modules:

When the head is attached to one or more disk shelves, ESH and LRC modules are the connection point between the head and disk shelves.

Consistency groups

Consistency groups makes sure that data is been replicated consistently
In other words data which is been written to set of lun will be replicated as point-in-time

Cache flushing techniques

Cache flushing methods

Idle cache flushing
Watermark cache flushing
Force cache flushing

If In general cache writes data to disk in small chunks  but the data which has to be written from the host end is very large then idle cache flushing cannot maintain LWM(low watermark level) in this situation watermark cache flushing kicks in

Watermark cache flushing try's to maintain cache between low and high watermark if It fails to bring cache below high watermark then force cache flushing kicks in disabling all write cache

Force cache flushing writes data to disk by disabling  all write cache till data in cache reaches LWM 
Only after then write cache is enabled

In the mean while data wot be written to cache and rather will be sent directly to disks resulting in low write performance

By default 
LWM - 20%
HWM - 80%

LUN Migration

Right Click on the source LUN and select migrate
and then select LUN dialog box appears.

Why LUN Migration ?

In order to overcome any of the these limitations LUNs are migrated

1. Speed
2. Performance
3. Capacity

LUN Assignment for two cluster nodes

Each cluster node can have single storage group

Create two storage groups assigned to same host
Add LUN to each storage group.

------------- OR ------------------

Create a storage group and map two nodes and LUN to that storage group.

Meta Creation

Base LUN + New LUN (Unassigned LUN)

Goto LUN Expand Wizard.

Choose the Base LUN.

Choose the unassigned LUN which has to be added to the Base LUN to expand.

Choose the type of meta.

Stripping - Takes time as the existing data has to be striped equally between two LUN's.
Concat - Just adds space on a fly.