WAFL Simplified

Module 5
Data ONTAP 7-Mode Administration
Module Objectives

By the end of this module, you should be able to:

 Describe how data is written to and read from
a WAFL® (Write Anywhere File Layout) file
system on a volume
 Explain the WAFL file system concepts,
including consistency points (CPs), RAID
management, and storage levels
 Describe how RAID is used to protect disk data
 Explain how the WAFL file system processes
write and read requests

© 2011 NetApp, Inc. All rights reserved. 2

Data ONTAP 7.3.x Architecture

Network Protocols WAFL RAID Storage

NVRAM
Clients Physical
Memory
Disk
Array

© 2011 NetApp, Inc. All rights reserved. 3

Data ONTAP 8.0.x 7-Mode Architecture

D-Blade

Client Protocol
Access Network Protocols WAFL RAID Storage
M-Host

NVRAM
Clients Physical
Memory
Disk
FreeBSD Array

© 2011 NetApp, Inc. All rights reserved. 4

Write Requests

© 2011 NetApp, Inc. All rights reserved. 5

Write Requests

 The Data ONTAP operating system receives write

requests through multiple protocols:
– CIFS
– NFS
– Fibre Channel (FC)
– iSCSI
– HTTP
– Web-based Distributed Authoring and Versioning
(WebDAV)
 Write requests are buffered into:
– System memory
– Nonvolatile RAM (NVRAM)
© 2011 NetApp, Inc. All rights reserved. 6
Write Request Data Flow: Write Buffer

Network
Network Protocols NVLOG N
Stack NVLOG V
Memory Buffer /
NVLOG R
SAN Cache
Service
NVLOG A
RS-232
NVLOG M
NVRAM Full
SAN Host NFS WAFL
HBA Service

CIFS
UNIX NIC Service
Client
RAID

Windows
Client
Storage

© 2011 NetApp, Inc. All rights reserved. 7

Consistency Point

 A CP is a completely self-consistent image of a file

system.
 A CP is equivalent to capturing the structure of a file
system at a moment in time.
 When a CP occurs, designated data is written to a disk
and a new root inode is determined.
 A CP occurs for multiple reasons, including but not
limited to the following:
– Half of the NVRAM card is full
– 10 seconds elapse
– A Snapshot copy is created (discussed in Module 12)
– The system is halted

© 2011 NetApp, Inc. All rights reserved. 8

CPs in the Data ONTAP Operating System

 For a CP, the Data ONTAP operating system

flushes writes to disk
– It always writes to new data blocks.
– The volume is always consistent on the disk.
 When the Data ONTAP operating system
flushes memory to disk:
– It updates the file system “atomically,” meaning
that the entire write must be completed or the
entire write is rolled back.
– The flush includes all metadata.
– It checks and then clears the NVRAM.

© 2011 NetApp, Inc. All rights reserved. 9

Write Request Data Flow: WAFL to RAID

Network
Network Protocols NVLOG N
Stack NVLOG V
Memory Buffer /
NVLOG R
SAN Cache
Service
NVLOG A
RS-232
NVLOG M
NVRAM Full
SAN Host NFS WAFL
HBA Service

CIFS
UNIX NIC Service
Client
RAID

Windows
Client
Storage

© 2011 NetApp, Inc. All rights reserved. 10

CPs from the WAFL File System to RAID
 The RAID layer calculates the parity of the data:
– To protect it from one or more disk failures
– To protect stripes of data
 The RAID layer calculates checksums, stored in
the block or zone method.
 If a data disk fails, the missing information can be
calculated from parity.
 The storage system can be configured in one of
two ways:
– RAID 4: The system can recover from one disk
failure in the RAID group.
– RAID-DP®: The system can recover from up to two
disk failures in the RAID group.

© 2011 NetApp, Inc. All rights reserved. 11

Write Request Data Flow: RAID to Storage

Network
Network Protocols NVLOG N
Stack NVLOG V
Memory Buffer /
NVLOG R
SAN Cache
Service
NVLOG A
RS-232
NVLOG M
NVRAM Full
SAN Host NFS WAFL
HBA Service

CIFS
UNIX NIC Service
Client
RAID
4k
Windows
Client Checksum
computed Storage

© 2011 NetApp, Inc. All rights reserved. 12

CPs from RAID to Storage

 The storage layer commits data and parity to

the physical disks.
 The root inode is updated to point to the new
file inodes on the disk.
 NVRAM is flushed and made available.
 The CP now is complete.

© 2011 NetApp, Inc. All rights reserved. 13

Write Request Data Flow: Storage Writes

Network
Network Protocols NVLOG N
Stack NVLOG V
Memory Buffer /
NVLOG R
SAN Cache
Service
NVLOG A
RS-232
NVLOG M
NVRAM Full
SAN Host NFS WAFL
HBA Service

CIFS
UNIX NIC Service
Client
RAID

Windows
Client
Storage

© 2011 NetApp, Inc. All rights reserved. 14

NVRAM

 The Data ONTAP operating system writes from

system memory:
– NVRAM is never read during normal write
operations.
– NVRAM is backed up with a battery.
 If a system failure occurs before the
completion of a CP, the data is read from
NVRAM and added back to the system
memory buffer when the system is brought
back online (or by the partner machine in a
high-availability controller configuration).

© 2011 NetApp, Inc. All rights reserved. 15

Read Requests

© 2011 NetApp, Inc. All rights reserved. 16

Read Requests

 Every time a read request is received, the

WAFL file system does one of two things:
– Reads the data from the system memory, also
known as the “cache”
– Reads the data from the disks
 The cache is populated by:
– Data recently read from disk
– Data recently written to disk

© 2011 NetApp, Inc. All rights reserved. 17

Read Request Data Flow: Read from Disk

Network
Network Protocols N
Stack V
Memory Buffer /
R
SAN Cache
Service
A
Console
M

SAN Host NFS WAFL

HBA Service

CIFS
UNIX NIC Service
Client
RAID

Windows
Client
Storage

© 2011 NetApp, Inc. All rights reserved. 18

Read Request Data Flow: Cache

Network
Network Protocols N
Stack V
Memory Buffer /
R
SAN Cache
Service
A
RS-232
M

SAN Host NFS WAFL

HBA Service

CIFS
UNIX NIC Service
Client
RAID

Windows
Client
Storage

© 2011 NetApp, Inc. All rights reserved. 19

Module Summary

In this module, you should have learned to:

 Describe how data is written to and read from
a WAFL file system on a volume
 Explain the WAFL file system concepts,
including CPs, RAID management, and
storage levels
 Describe how RAID is used to protect disk data
 Explain how the WAFL file system processes
write and read requests

© 2011 NetApp, Inc. All rights reserved. 20

Exercise
Module 5: WAFL Simplified
Estimated Time: 15 minutes
Check Your Understanding: Answers

 What is a consistency point?

A completely self-consistent image of the entire
file system
 What is the purpose of RAID?
RAID is a method used to protect disk data by
which a parity value is calculated across the disks.
 What does the storage layer do?
Handles the I/O operations to disk and checks for
disk failures

© 2011 NetApp, Inc. All rights reserved. 24