VMware | NetApp FAS 2050HA | Specifications | VMware NetApp FAS 2050HA Specifications

NETAPP TECHNICAL REPORT
NetApp Deduplication for FAS and V-Series
Deployment and Implementation Guide
Carlos Alvarez
March 2009 | TR-3505-0309
ABSTRACT
®
This guide introduces NetApp deduplication for FAS technology, describes in detail how to implement and
use it, and provides information on best practices, operational considerations, and troubleshooting.
It should prove useful for both NetApp and channel partner sales and services field personnel who require
assistance in understanding details and successfully deploying solutions that include deduplication.
TABLE OF CONTENTS
1
2
3
4
2
INTRODUCTION AND OVERVIEW OF DEDUPLICATION ....................................................... 4
1.1
HOW DEDUPLICATION FOR FAS WORKS ............................................................................................... 4
1.2
DEDUPLICATED VOLUMES ....................................................................................................................... 5
1.3
DEDUPLICATION METADATA ................................................................................................................... 5
1.4
GENERAL DEDUPLICATION FEATURES .................................................................................................. 7
CONFIGURATION AND OPERATION ....................................................................................... 7
2.1
REQUIREMENTS OVERVIEW .................................................................................................................... 7
2.2
INSTALLING AND LICENSING DEDUPLICATION ..................................................................................... 8
2.3
COMMAND SUMMARY ............................................................................................................................... 9
2.4
DEDUPLICATION QUICK START GUIDE................................................................................................. 10
2.5
END-TO-END DEDUPLICATION CONFIGURATION EXAMPLE .............................................................. 10
2.6
CONFIGURING DEDUPLICATION SCHEDULES ..................................................................................... 12
SIZING FOR PERFORMANCE AND SPACE EFFICIENCY .................................................... 13
3.1
DEDUPLICATION GENERAL BEST PRACTICES .................................................................................... 13
3.2
DEDUPLICATION PERFORMANCE ......................................................................................................... 14
3.3
DEDUPLICATION STORAGE SAVINGS................................................................................................... 16
3.4
SPACE SAVINGS ESTIMATION TOOL (SSET) ........................................................................................ 17
3.5
DEDUPLICATION LIMITATIONS .............................................................................................................. 18
DEDUPLICATION WITH OTHER NETAPP FEATURES ......................................................... 20
4.1
DEDUPLICATION AND SNAPSHOT COPIES .......................................................................................... 20
4.2
DEDUPLICATION AND SNAPRESTORE ................................................................................................. 21
4.3
DEDUPLICATION AND THE VOL COPY COMMAND .............................................................................. 21
4.4
DEDUPLICATION AND READ REALLOCATION (REALLOC) ................................................................. 21
4.5
DEDUPLICATION AND FLEXCLONE VOLUMES .................................................................................... 21
4.6
DEDUPLICATION AND ACTIVE-ACTIVE CONFIGURATION................................................................... 22
4.7
DEDUPLICATION AND V-SERIES ............................................................................................................ 22
4.8
DEDUPLICATION AND SNAPMIRROR REPLICATION ........................................................................... 23
4.9
DEDUPLICATION AND SNAPVAULT....................................................................................................... 27
4.10
DEDUPLICATION AND SNAPVAULT FOR NETBACKUP ....................................................................... 27
4.11
DEDUPLICATION AND MULTISTORE (VFILER) ..................................................................................... 28
4.12
DEDUPLICATION AND SNAPLOCK ........................................................................................................ 28
4.13
DEDUPLICATION AND METROCLUSTER ............................................................................................... 28
4.14
DEDUPLICATION AND DATAFORT ENCRYPTION ................................................................................. 29
4.15
DEDUPLICATION AND LUNS................................................................................................................... 29
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
5
DEDUPLICATION AND VMWARE ........................................................................................... 33
5.1
VMFS DATA STORE ON FIBRE CHANNEL OR ISCSI: SINGLE LUN ..................................................... 33
5.2
VMWARE VIRTUAL DISKS OVER NFS/CIFS ........................................................................................... 34
5.3
DEDUPLICATION ARCHIVE OF VMWARE .............................................................................................. 35
6
DEDUPLICATION AND SHAREPOINT ................................................................................... 35
7
DEDUPLICATION AND EXCHANGE ....................................................................................... 36
8
DEDUPLICATION AND TIVOLI STORAGE MANAGER (TSM) .............................................. 36
9
DEDUPLICATION AND SYMANTEC BACKUP EXEC ........................................................... 36
10 DEDUPLICATION AND LOTUS DOMINO ............................................................................... 36
11 TROUBLESHOOTING .............................................................................................................. 37
11.1
LICENSING ............................................................................................................................................... 37
11.2
VOLUME SIZES ........................................................................................................................................ 37
11.3
LOGS AND ERROR MESSAGES ............................................................................................................. 37
11.4
NOT SEEING SPACE SAVINGS ............................................................................................................... 37
11.5
UNDEDUPLICATING A FLEXIBLE VOLUME ........................................................................................... 38
11.6
ADDITIONAL REPORTING WITH SIS STAT -1 ........................................................................................ 39
12 ADDITIONAL READING AND REFERENCES ........................................................................ 39
13 ADDITIONAL ASSISTANCE .................................................................................................... 40
3
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
1
INTRODUCTION AND OVERVIEW OF DEDUPLICATION
This section provides an overview of how deduplication works for FAS and V-Series systems.
Notes:
1.
Whenever references are made to deduplication for FAS in this document, the reader should assume
that the same information also applies to V-Series systems, unless otherwise noted.
2.
NetApp deduplication for VTL is not covered within this technical report and is beyond the scope of this
document.
1.1
HOW DEDUPLICATION FOR FAS WORKS
Part of NetApp’s storage efficiency offerings, NetApp deduplication for FAS provides block-level
deduplication within the entire flexible volume on NetApp storage systems. Beginning with Data ONTAP®
7.3, V-Series also supports deduplication. NetApp V-Series is designed to be used as a gateway system
that sits in front of third-party storage, allowing NetApp storage efficiency and other features to be used on
third-party storage.
Figure 1 shows how NetApp deduplication for FAS works at the highest level.
Figure 1) How NetApp deduplication for FAS works.
Essentially, deduplication stores only unique blocks in the flexible volume and creates a small amount of
additional metadata in the process. Notable features of deduplication include:
It works with a high degree of granularity; that is, at the 4KB block level.
It operates on the active file system of the flexible volume. Any block referenced by a Snapshot ™ copy
is not made ―available‖ until the Snapshot copy is deleted.
It’s a background process that can be configured to run automatically, or it can be scheduled, or run
manually through the command line interface (CLI).
It’s application transparent, and therefore it can be used for deduplication of data originating from any
application using the NetApp system.
It’s enabled and managed using a simple CLI only at this time.
It can be enabled on and can deduplicate blocks on flexible volumes with new and existing data.
4
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
In summary, this is how deduplication works. Newly saved data on the FAS system is stored in 4KB blocks
as usual by Data ONTAP. Each block of data has a digital fingerprint, which is compared to all other
fingerprints in the flexible volume. If two fingerprints are found to be the same, a byte-for-byte comparison is
done of all bytes in the block and, if there is an exact match between the new block and the existing block on
the flexible volume, the duplicate block is discarded and its disk space is reclaimed.
1.2
DEDUPLICATED VOLUMES
Despite the introduction of less expensive ATA disk drives, one of the biggest challenges for storage
systems today continues to be the storage cost. There is a desire to reduce storage consumption (and
therefore storage cost per MB) by eliminating duplicate data through sharing blocks across files.
The core NetApp technology to accomplish this goal is the deduplicated volume, a flexible volume that
contains shared data blocks. Data ONTAP supports shared blocks in order to optimize storage space
consumption. Basically, within one volume, there is the ability to have multiple references to the same data
block, as shown in Figure 2.
Figure 2) Data structure in a deduplicated volume.
In Figure 2, the number of physical blocks used on the disk is 3 (instead of 5), and the number of blocks
saved by deduplication is 2 (5 minus 3). In the remainder of this document, these will be referred to as used
blocks and saved blocks.
Each data block has a block count reference kept in the volume metadata. As additional indirect blocks
(―IND‖ in Figure 2) point to the data, or existing ones stop pointing to it, this value is incremented or
decremented accordingly. When no indirect blocks point to a data block, it is released.
The NetApp deduplication technology allows duplicate 4KB blocks anywhere in the flexible volume to be
deleted, as described in the following sections.
The maximum sharing for a block is 255. This means, for example, that if there are 500 duplicate blocks,
deduplication would reduce that to only 2 blocks. Also note that this ability to share blocks is different from
the ability to keep 255 Snapshot copies for a volume.
1.3
DEDUPLICATION METADATA
The core enabling technology of deduplication is fingerprints. These are unique digital ―signatures‖ for every
4KB data block in the flexible volume.
When deduplication runs for the first time on a flexible volume with existing data, it scans the blocks in the
flexible volume and creates a fingerprint database, which contains a sorted list of all fingerprints for used
blocks in the flexible volume.
After the fingerprint file is created, fingerprints are checked for duplicates and, when found, first a byte-bybyte comparison of the blocks is done to make sure that the blocks are indeed identical, and if they are
found to be identical, the block’s pointer is updated to the already existing data block and the new (duplicate)
data block is released.
5
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
Releasing a duplicate data block entails updating the indirect inode pointing to it, incrementing the block
reference count for the already existing data block, and freeing the duplicate data block.
In real time, as additional data is written to the deduplicated volume, a fingerprint is created for each new
block and written to a change log file. When deduplication is run subsequently, the change log is sorted and
its sorted fingerprints are merged with those in the fingerprint file, and then the deduplication processing
occurs.
Note that there are really two change log files, so that as deduplication is running and merging the new
blocks from one change log file into the fingerprint file, new data that is being written to the flexible volume is
causing fingerprints for these new blocks to be written to the second change log file. The roles of the two
files are then reversed the next time that deduplication is run. (For those familiar with Data ONTAP usage of
NVRAM, this is analogous to when it switches from one half to the other to create a consistency point.)
Note: When deduplication is run for the first time on an empty flexible volume, it still creates the fingerprint
file from the change log.
Here are some additional details about the deduplication metadata:
There is a fingerprint record for every 4KB data block, and the fingerprints for all the data blocks in the
volume are stored in the fingerprint database file.
Fingerprints are not deleted from the fingerprint file automatically when data blocks are freed, but when
a threshold of 20% new fingerprints is reached, the stale fingerprints are deleted. This can also be done
by a manual operation from the command line.
In Data ONTAP 7.2.X, all the deduplication metadata resides in the flexible volume.
Starting with Data ONTAP 7.3.0, part of the metadata resides in the volume and part of it resides in the
aggregate outside the volume. The fingerprint database and the change log files that are used in the
deduplication process are located outside of the volume in the aggregate and are therefore not captured
in Snapshot copies. This change enables deduplication to achieve higher space savings. However,
some other temporary metadata files created during the deduplication operation are still placed inside
the volume. These temporary metadata files are deleted once the deduplication operation is complete.
These temporary metadata files can get locked in Snapshot copies if the Snapshot copies are created
during a deduplication operation. The metadata files remain locked until the Snapshot copies are
deleted.
During an upgrade from Data ONTAP 7.2 to 7.3, the fingerprint and change log files will be moved from
the flexible volume to the aggregate level during the next deduplication process following the upgrade.
During the deduplication process where the fingerprint and change log files are being moved from the
volume to the aggregate, the ―sis status‖ command will display the message ―Fingerprint is being
upgraded.‖
In Data ONTAP 7.3 and later, the deduplication metadata for a volume is located outside the volume, in
the aggregate. When you revert from Data ONTAP 7.3 to a pre-7.3 release, the deduplication metadata
is lost during the revert process. In order to obtain optimal space savings, use the sis start –s
command to rebuild the deduplication metadata for all existing data. If this is not done, the existing data
in the volume will retain the space savings from deduplication run prior to the revert process; however,
any deduplication that occurs after the revert process will only apply to data that was created after the
revert process, and will not deduplicate against data that existed prior to the revert process. The sis
start –s command can take a long time to complete, depending on the size of the logical data in the
volume, but during this time the system is available for all other operations. Before using the sis
start –s command, make sure that the volume has sufficient free space to accommodate the addition
of the deduplication metadata to the volume. The deduplication metadata uses 1% to 6% of the logical
data size in the volume.
For the size of the overhead associated with the deduplication metadata files, see the section
―Deduplication Metadata Overhead.‖
6
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
1.4
GENERAL DEDUPLICATION FEATURES
Deduplication is enabled on a per flexible volume basis. It can be enabled on any number of flexible
volumes in a storage system. It can be run one of four different ways:
Scheduled on specific days and at specific times
Manually using the command line
Automatically, when 20% new data has been written to the volume
Automatically on the destination volume, when used with SnapVault®
Only one deduplication process can run on a flexible volume at a time.
Up to eight deduplication processes can run concurrently on eight volumes within the same NetApp storage
system.
Beginning with Data ONTAP 7.3.1, deduplication checkpoint restart will allow a deduplication process that
was interrupted to continue from the last checkpoint. Prior to Data ONTAP 7.3.1, an interrupted
deduplication process would result in a restart of the entire deduplication process. If the system is restarted
while deduplication is in process, when the system is once again online, the deduplication process will
automatically restart from the last checkpoint.
2
CONFIGURATION AND OPERATION
This section discusses what is required to install deduplication, how to configure it, and various aspects of
managing it. Although it discusses some basic things, in general it assumes both that the NetApp storage
system is already installed and running and that the reader is familiar with basic NetApp administration.
2.1
REQUIREMENTS OVERVIEW
Table 1 specifies the hardware and software required to run deduplication.
Table 1) Deduplication requirements overview.
Hardware
NearStore® R200
FAS2000 series
FAS3000 series
FAS3100 series
FAS6000 series
IBM N5000 series
IBM N7000 series
Note: Starting with Data ONTAP 7.3, the V-Series systems corresponding to
the NetApp FAS systems and IBM N-Series Gateway systems listed above are
also supported.
Data ONTAP
Required minimum release is Data ONTAP 7.2.5.1 or later
Software
- nearstore_option license (for all platforms except R200)
- a_sis license
7
Maximum deduplication
volume sizes for different
data ONTAP versions
See section ‖Maximum Flexible Volume Size‖
Supported protocols
All
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
Here are some additional considerations with regard to max volume sizes for deduplication:
Once an upgrade to Data ONTAP 7.3.1 is complete, the new maximum volume sizes for Data ONTAP
7.3.1 will be in effect.
If considering a downgrade or revert, it is highly recommended that NetApp Global Services be
consulted for best practice.
During a revert from Data ONTAP 7.3.1 to an earlier version of Data ONTAP with smaller volume limits,
volumes should be within the limits of the lower version of Data ONTAP. If downgrading to 7.2.5.1, you
cannot simply resize the volume. Instead, a new flexible volume that is within the max vol size limits
must be created, and the data must be moved to that volume prior to the downgrade.
If a downgrade has occurred from 7.3.1 to 7.3.0, where the volume size was greater than the 7.3.0
volume size limit, then the volume will go offline. If this has occurred, contact NetApp Global Services
for assistance with bringing the volume back online.
WHAT’S SUPPORTED WITH DEDUPLICATION
The following NetApp features are supported with deduplication:
Deduplication is supported on the R200 systems and on all FAS and V-Series systems with the
NearStore option license.
Only flexible volumes are supported. Traditional volumes are not supported.
Starting with Data ONTAP 7.3.1, SnapLock® volumes are supported with deduplication in both
enterprise and compliance modes.
LUNs are supported with deduplication.
SnapMirror® is supported with deduplication (both qtree SnapMirror and volume SnapMirror).
SnapMirror sync mode is not supported with deduplication.
SnapVault on the source volume is supported with deduplication.
Starting with Data ONTAP 7.3, SnapVault on the destination volume is supported.
Starting with Data ONTAP 7.3, Open Systems SnapVault is supported.
Starting with Data ONTAP 7.3.1, stretch MetroCluster is supported with deduplication.
Starting with Data ONTAP 7.3.1, fabric MetroCluster is supported with deduplication.
Starting with Data ONTAP 7.3, the V-Series product line is supported with deduplication.
Starting with Data ONTAP 7.3, MultiStore® is supported with deduplication. Starting with Data ONTAP
7.3.1, deduplication commands are also available from within each MutiStore vFiler™ unit.
FlexShare™ is supported with deduplication.
NDMP dump is supported with deduplication.
2.2
INSTALLING AND LICENSING DEDUPLICATION
Deduplication is included in Data ONTAP and just needs to be licensed. Add the deduplication license by
using the following command:
license add <a_sis license key>
To run deduplication on any of the FAS platforms, you also need to add the NearStore option license:
license add <nearstore_option license key>
There is no charge for either license; contact your sales representative to obtain the licenses.
DEDUPLICATION LICENSING IN A CLUSTERED ENVIRONMENT
Deduplication is a licensed option behind the NearStore option license. Both nodes must have the
NearStore option licensed. Deduplication must be licensed on both nodes of the cluster as well.
8
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
2.3
COMMAND SUMMARY
Table 2 describes all deduplication (related) commands.
Table 2) Deduplication command summary.
Command
Summary
sis on <vol>
Enables deduplication on the specified flexible volume.
sis start -s <vol>
Begins the deduplication process on the flexible volume specified and performs
a scan of the flexible volume to process existing data.
This option is typically used upon initial configuration and deduplication on an
existing flexible volume that contains undeduplicated data. (There’s no need to
use this option on a volume that has just been created and doesn’t contain any
data.)
sis start -sp <vol>
Begins the deduplication process on the flexible volume specified using the
existing checkpoint information, regardless of the age of the checkpoint
information.
This option should only be used with -s.
sis start -d <vol>
Deletes the existing checkpoint information.
This option is used to delete checkpoint information that is still considered
valid. By default, checkpoint information is considered invalid after 24 hours.
sis start <vol>
Begins the deduplication process on the flexible volume specified.
sis status [-l] <vol>
Returns the current status of deduplication for the specified flexible volume.
The -l option displays a long list.
df –s <vol>
Returns the value of deduplication space savings in the active file system for
the specified flexible volume. Use this command to see how much space has
been saved.
sis config [-s sched]\ <vol>
Creates an automated deduplication schedule.
When deduplication is first enabled on a flexible volume, a default schedule is
configured, running it each day of the week at midnight. If the auto option is
used, deduplication is triggered when 20% new data is written to the volume.
Starting with 7.3.1, the 20% threshold can be adjusted using the auto@num
option, where num is a two-digit number to specify the percentage.
sis stop <vol>
Suspends an active deduplication process on the flexible volume specified.
sis off <vol>
Deactivates deduplication on the flexible volume specified. This means that
there will be no more change logging or deduplication operations, but the
flexible volume remains a deduplicated volume and the storage savings are
kept.
If this command is used, and then deduplication is turned back on for this
flexible volume, the flexible volume must be rescanned with the sis start –
s command.
9
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
Command
Summary
sis check <vol>
(This command is available
only in Diag mode.)
Verifies and updates the fingerprint database for the specified flexible volume;
includes purging stale fingerprints.
sis stat <vol>
(This command is available
only in Diag mode.)
Displays the statistics of flexible volumes that have deduplication enabled.
sis undo <vol>
(This command is available
in Advanced and Diag
modes.)
Reverts a deduplicated volume to a normal flexible volume.
2.4
DEDUPLICATION QUICK START GUIDE
This section provides a quick run-through of the steps to configure and manage deduplication.
Table 3) Deduplication quick overview.
New Flexible Volume
Flexible Volume with Existing Data
Flexible volume configuration
Create flexible volume.
N/A
Enable deduplication on flexible
volume
sis on <vol>
Initial scan
N/A
Scan and deduplicate the existing data.
sis start -s <vol>
Create, modify, delete
schedules (if not doing
manually)
Delete or modify the default deduplication schedule that was configured
when deduplication was first enabled on the flexible volume, or create the
desired schedule.
sis config [-s sched] <vol>
Manually run deduplication (if
not using schedules)
sis start <vol>
Monitor status of deduplication
(optional)
sis status <vol>
Monitor space savings
(optional)
df –s <vol>
2.5
END-TO-END DEDUPLICATION CONFIGURATION EXAMPLE
This section steps through the entire process of creating a flexible volume and configuring, running, and
monitoring deduplication on it.
Note: The steps are spelled out in detail, so the process appears a lot lengthier than it would be in the real
world.
This example creates a place to archive several large data files. The destination NetApp storage system is
called r200-rtp01, and it is assumed that deduplication has been licensed on this machine.
10
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
1.
Create a flexible volume (keeping in mind the maximum allowable volume size for the platform, as
specified in the requirements table at the beginning of this section).
r200-rtp01*> vol create VolArchive aggr0 200g
Creation of volume 'VolArchive' with size 200g on containing aggregate
'aggr0' has completed.
2.
Enable deduplication on the flexible volume and verify that it’s turned on. The vol status command
shows the attributes for flexible volumes that have deduplication turned on.
After you turn deduplication on, Data ONTAP lets you know that if this were an existing flexible volume
that already contained data before deduplication was enabled, you would want to run sis start –s.
In this example it’s a brand-new flexible volume, so that’s not necessary.
r200-rtp01> sis on /vol/VolArchive
Deduplication for "/vol/VolArchive" is enabled.
Already existing data could be processed by running "sis start -s
/vol/VolArchive.”
r200-rtp01> vol status VolArchive
Volume State
Status
VolArchive online raid_dp, flex
sis
Containing aggregate: 'aggr0'
3.
Options
nosnap=on
Another way to verify that deduplication is enabled on the flexible volume is to check the output from
running sis status on the flexible volume.
r200-rtp01> sis status /vol/VolArchive
Path
State
Status
/vol/VolArchive
Enabled
Idle
4.
Progress
Idle for 00:00:20
Turn off the default deduplication schedule.
r200-rtp01> sis config /vol/VolArchive
Path
Schedule
/vol/VolArchive
sun-sat@0
r200-rtp01> sis config -s - /vol/VolArchive
r200-rtp01> sis config /vol/VolArchive
Path
Schedule
/vol/VolArchive
5.
-
NFS-mount the flexible volume to /testArchives on a Solaris™ host called sunv240-rtp01, and
copy lots of files from the users’ directories into the new archive directory flexible volume. Here is the
result from the host perspective:
root@sunv240-rtp01 # pwd
/testPSTs
root@sunv240-rtp01 # df -k.
Filesystem
kbytes
used
avail capacity Mounted on
r200-rtp01:/vol/VolArchive 167772160 33388384 134383776 20%
/testArchives
6.
Next, examine the flexible volume, run deduplication, and monitor the status. Use the df –s command
to examine the storage consumed and the space saved. Note that no space savings have been
achieved by simply copying data to the flexible volume, even though deduplication is turned on. What
has happened is that all the blocks that have been written to this flexible volume since deduplication
was turned on have had their fingerprints written to the change log file.
r200-rtp01> df -s /vol/VolArchive
Filesystem
used
/vol/VolArchive/
33388384
11
saved
0
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
%saved
0%
7.
Run deduplication on the flexible volume. These causes the change log to be processed, fingerprints to
be sorted and merged, and duplicate blocks to be found.
r200-rtp01> sis start /vol/VolArchive
The deduplication operation for "/vol/VolArchive" is started.
8.
9.
Use sis status to monitor the progress of deduplication.
r200-rtp01> sis status /vol/VolArchive
Path
State
Status
/vol/VolArchive
Enabled
Active
Progress
9211 MB Searched
r200-rtp01> sis status /vol/VolArchive
Path
State
Status
/vol/VolArchive
Enabled
Active
Progress
11 MB (0%) Done
r200-rtp01> sis status /vol/VolArchive
Path
State
Status
/vol/VolArchive
Enabled
Active
Progress
1692 MB (14%) Done
r200-rtp01> sis status /vol/VolArchive
Path
State
Status
/vol/VolArchive
Enabled
Active
Progress
10 GB (90%) Done
r200-rtp01> sis status /vol/VolArchive
Path
State
Status
/vol/VolArchive
Enabled
Active
Progress
11 GB (99%) Done
r200-rtp01> sis status /vol/VolArchive
Path
State
Status
/vol/VolArchive
Enabled
Idle
Progress
for 00:00:07
When sis status indicates that the flexible volume is once again in the Idle state, deduplication has
finished running, and you can check the space savings it provided in the flexible volume.
r200-rtp01> df -s /vol/VolArchive
Filesystem
used
/vol/VolArchive/
24072140
saved
9316052
%saved
28%
That’s all there is to it.
2.6
CONFIGURING DEDUPLICATION SCHEDULES
It’s best to set up a schedule for deduplication so that you don’t have to run it manually each time. This
section provides some specifics about configuring schedules with deduplication.
The sis config command is used to configure and view deduplication schedules for flexible volumes. The
usage syntax is shown below.
r200-rtp01> sis help config
sis config [ [ -s schedule ] <path> | <path> ... ]
Sets up, modifies, and retrieves the schedule of deduplication volumes.
Run with no arguments, sis config returns the schedules for all flexible volumes that have deduplication
enabled. The following example shows the four different formats the reported schedules can have.
toaster> sis config
Path
/vol/dvol_1
/vol/dvol_2
/vol/dvol_3
/vol/dvol_4
Schedule
23@sun-fri
auto
sat@6
The meaning of each of these schedule types is as follows:
On flexible volume dvol_1, deduplication is not scheduled to run.
12
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
On flexible volume dvol_2, deduplication is scheduled to run every day from Sunday to Friday at 11
p.m.
On flexible volume dvol_3, deduplication is set to autoschedule. This means that deduplication is
triggered by the amount of new data written to the flexible volume, specifically when there are 20% new
fingerprints in the change log.
On flexible volume dvol_4, deduplication is scheduled to run at 6 a.m. on Saturday.
When the -s option is specified, the command sets up or modifies the schedule on the specified flexible
volume. The schedule parameter can be specified in one of four ways:
[day_list][@hour_list]
[hour_list][@day_list]
auto
The day_list specifies which days of the week deduplication should run. It is a comma-separated list of
the first three letters of the day: sun, mon, tue, wed, thu, fri, sat. The names are not case sensitive. Day
ranges such as mon-fri can also be used. The default day_list is sun-sat.
The hour_list specifies which hours of the day deduplication should run on each scheduled day. The
hour_list is a comma-separated list of the integers from 0 to 23. Hour ranges such as 8-17 are allowed.
Step values can be used in conjunction with ranges. For example, 0-23/2 means "every 2 hours." The
default hour_list is 0; that is, midnight on the morning of each scheduled day.
If "-" is specified, there is no scheduled deduplication operation on the flexible volume.
The auto schedule causes deduplication to run on that flexible volume whenever there are 20% new
fingerprints in the change log. This check is done in a background process and occurs every hour.
When deduplication is enabled on a flexible volume for the first time, an initial schedule is assigned to the
flexible volume. This initial schedule is sun-sat@0, which means "once every day at midnight."
To configure the schedules shown earlier in this section, the following commands would be issued:
toaster>
toaster>
toaster>
toaster>
3
sis
sis
sis
sis
config
config
config
config
-s
-s
–s
–s
- /vol/dvol_1
23@sun-fri /vol/dvol_2
auto /vol/dvol3
sat@6 /vol/dvol_4
SIZING FOR PERFORMANCE AND SPACE EFFICIENCY
This section discusses the deduplication behavior that you can expect. Information in this section comes
from testing, observations, and knowledge of how deduplication functions.
3.1
DEDUPLICATION GENERAL BEST PRACTICES
This section contains deduplication best practices and lessons learned from internal tests and from
deployments in the field.
Deduplication consumes system resources and can alter the data layout on disk. Due to the
application’s I/O pattern and the effect of deduplication on the data layout, the read and write I/O
performance can vary considerably. The space savings and the performance impact vary significantly
depending on the application and the data contents.
NetApp recommends that the performance impact due to deduplication be carefully considered and
measured in a test setup and taken into sizing considerations before deploying deduplication in
performance-sensitive solutions. For more information on the impact of deduplication on other
applications, contact the specialists at NetApp for their advice and test results of your particular
application with deduplication.
If there is a small amount of new data, run deduplication infrequently, because there’s no benefit in
running it frequently in such a case, and it consumes CPU resources. How often you run it depends on
the rate of change of the data in the flexible volume.
13
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
The more concurrent deduplication processes you’re running, the more system resources are
consumed.
Given the previous two items, the best option is to do one of the following:
Use the auto mode so that deduplication runs only when significant additional data has been written to
each particular flexible volume (this tends to naturally spread out when deduplication runs).
Stagger the deduplication schedule for the flexible volumes so that it runs on alternative days.
Run deduplication manually.
If Snapshot copies are required, run deduplication before creating the Snapshot copy to minimize the
amount of data before the data gets locked in to the copies. (Make sure that deduplication has
completed before creating the copy.) If a Snapshot copy is created on a flexible volume before
deduplication has a chance to complete on that flexible volume, this could result in lower space savings.
If Snapshot copies are to be used, the Snapshot reserve should be greater than 0. An exception to this
could be in a volume that contains LUNs, where snap reserve might be set to zero for thin provisioning
reasons, and additional free space should be available in the volume to contain Snapshot copies.
For deduplication to run properly, you need to leave some free space for the deduplication metadata.
For information about how much extra space to leave in the volume and in the aggregate, see section
―Deduplication Metadata Overhead.‖
3.2
DEDUPLICATION PERFORMANCE
This section discusses the performance aspects of deduplication.
®
Since deduplication is a part of Data ONTAP, it is tightly integrated with the WAFL file structure.
Because of this, deduplication is performed with high efficiency. It is able to leverage the internal
characteristics of Data ONTAP to create and compare digital fingerprints, redirect data pointers, and
free up redundant data areas.
However, the following factors can affect the performance of the deduplication process and the I/O
performance of deduplicated volumes.
The application and the type of data set being used
The data access pattern (for example, sequential vs. random access, the size and pattern of the I/O)
The amount of duplicate data, the amount of total data, and the average file size
The nature of the data layout in the volume
The amount of changed data between deduplication runs
The number of concurrent deduplication sessions
Hardware platform—the amount of CPU/memory in the system
Amount of load on the system
Disk types ATA/FC, and the RPM of the disk
Number of disk spindles in the aggregate
Because of these factors, NetApp recommends that the performance impact due to deduplication be
carefully considered and measured in a test setup and taken into sizing considerations
before deploying deduplication in performance-sensitive solutions.
THE PERFORMANCE OF THE DEDUPLICATION OPERATION
The performance of the deduplication operation itself varies widely depending on the factors listed above,
and this determines how long it takes this background process to finish running.
On a FAS6080 with no other load on the system, we have seen deduplication performances of up to 120
MBytes/sec (running a single deduplication session). If multiple deduplication streams are running, this total
bandwidth gets divided evenly into the number of streams.
To get an idea of how long it takes for a deduplication process to complete, let’s say that the deduplication
process is running on a flexible volume at 25MB/sec. If 1TB of new data has been added to the volume
since the last deduplication update, this deduplication operation takes about 10 to 12 hours to complete.
14
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
(There are no configurable parameters that can tune the deduplication process; that is, the priority of this
background process in Data ONTAP is fixed.)
IMPACT ON THE SYSTEM DURING THE DEDUPLICATION PROCESS
The deduplication operation runs as a low-priority background process on the system. However, it can still
affect the performance of user I/O and other applications running on the system.
The number of deduplication processes that are running and the phase that each process is running in can
cause performance impacts to other applications running on the system (up to eight deduplication processes
can actively run at any time on a system). The following are some observations made when running
deduplication on a FAS3050 system:
With eight deduplication processes running, and no other processes running, deduplication uses 15% of
the CPU in its least invasive phase, and nearly all the available CPU in its most invasive phase.
When one deduplication process is running, there is a 0% to 15% performance degradation on other
applications.
With eight deduplication processes running, there may be as much as a 15% to 50% performance
penalty on other applications running on the system.
THE I/O PERFORMANCE OF DEDUPLICATED VOLUMES
Write Performance to a Deduplicated Volume
The impact of deduplication on the write performance of a system is a function of the hardware platform that
is being used, as well as the amount of load that is placed on the system.
If the load on a system is low—that is, for systems in which the CPU utilization is around 50% or lower—
there is a negligible difference in performance when writing data to a deduplicated volume, and there is no
noticeable impact on other applications running on the system. On heavily used systems, however, where
the system is nearly saturated with the amount of load on it, the impact on write performance can be
expected to be around 15% for most NetApp systems. The performance impact is more noticeable on
higher-end systems than on lower-end systems. On the FAS6080 system, this performance impact can be
as much as 35%. The higher degradation is usually experienced in association with random writes. Note that
these numbers are for FC drives; if ATA drives are used in a system, the performance impact would be
greater.
Read Performance from a Deduplicated Volume
When data is read from a deduplication-enabled volume, the impact on the read performance varies
depending on the difference between the deduplicated block layout compared to the original block layout.
There is minimal impact on random reads.
Since deduplication alters the data layout on the disk, it can affect the performance of sequential read
applications such as dump source, qtree SnapMirror or SnapVault source, SnapVault restore, and other
sequential read-heavy applications. This impact is more noticeable with data sets that are zero-padded or
data sets that contain blocks with repeating patterns (such as applications that preinitialize data blocks to a
value of zero). Significant performance degradation has been measured in sequential reads with these types
of data patterns. The performance impact is also more significant on sequential reads from SATA drives as
compared to FC drives. Therefore, if an application depends on sequential read performance, the impact of
deduplication on read performance should be carefully considered before implementation.
In cases of random reads on deduplicated volumes, there is usually little impact on performance.
THE PERFORMANCE ACCELERATION MODULE (PAM)
The PAM card is available with Data ONTAP 7.3 and later.
In environments where there are shared blocks that are read repeatedly, the PAM card can help reduce the
number of disk reads, thus improving the read performance.
The amount of performance improvement with the PAM card depends on the duplication rate, the access
rate, the active data set size, and the data layout.
Adding a PAM card to a system does not increase the deduplication maximum volume size for that system.
15
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
The PAM card has provided significant performance improvements in VMware ® VDI environments. The
advantages provided by the NetApp PAM are further enhanced when combined with other shared block
technologies such as NetApp deduplication or FlexClone ®.
For additional information regarding the PAM card, refer to TR-3705, NetApp and VMware VDI Best
Practices.
3.3
DEDUPLICATION STORAGE SAVINGS
This section discusses storage savings that deduplication can be expected to deliver.
Comprehensive testing of various data sets has been performed to determine typical space savings in
different environments. These results were obtained in three ways:
1.
Running deduplication on various production data sets within NetApp.
2.
NetApp systems deployed in the real world running deduplication.
3.
NetApp and end users running a simulation tool on various data sets. See section 3.4, ―Space Savings
Estimation Tool,‖ for information about how to use this tool.
Table 4 summarizes the test results.
Table 4) Typical deduplication storage savings for various environments.
Data Types
Typical Space Savings
Backup data
95%
VMware
70%
Hyper-V
60%
File Services
35%
Sharepoint
30%
Email Archival
30%
Document Archival
25%
Source Code Archival
25%
Audio/Video Files
10%
Note that nonrepeating archival data such as image files and encrypted data is generally not considered a
good candidate for deduplication.
The results reported in Table 4 are considered realistic and typically achievable, but still conservative.
Results can be validated in an existing environment using the Space Savings Estimation Tool, as discussed
in section 3.4.
It is important to note that the space savings in the table are from deduplicating a data set one time, with the
following exception. In cases where the data is being backed up or archived over and over again, the
realized storage savings get better and better, achieving 20:1 (95%) in many instances.
DEDUPLICATION AND SPACE SAVINGS ON EXISTING DATA
A major benefit of deduplication is that it can be used to deduplicate existing data in the flexible volumes. It
is realistic to assume that there will be Snapshot copies—perhaps many—of this existing data. Here’s what
happens when you run deduplication in this case.
When you first run deduplication on this flexible volume, the storage savings will probably be rather small or
even nonexistent.
Previous Snapshot copies expire, and as they do some small savings are realized, but they too are
likely to be low.
During this period of old Snapshot copies expiring, it is fair to assume that new data is being created on
the flexible volume and that Snapshot copies are being created.
16
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
The storage savings may continue to stay low.
When the last Snapshot copy that was created before deduplication was run is deleted, the storage
savings should increase noticeably.
The question thus becomes when to run deduplication again in order to achieve maximum capacity savings.
The answer is that deduplication should be run, and allowed to complete, before the creation of each and
every Snapshot copy; this provides the most storage savings benefit. However, depending on the flexible
volume size and possible performance impact on the system, this may not always be advisable.
DEDUPLICATION METADATA OVERHEAD
This section discusses storage overhead that deduplication introduces. While deduplication can provide
substantial storage savings in many environments, there is a small amount of storage overhead associated
with it. This should be considered when sizing the flexible volume.
The total storage used by the deduplication metadata files is approximately 1% to 6% of the total data in the
volume. Total data = used space + saved s pace, as reported when using df –s (that is, the size of the
data before it is deduplicated). So for 1TB of total data, the metadata overhead would be approximately
10GB to 60GB. The breakdown of the overhead associated with the deduplication metadata is as follows:
There is a fingerprint record for every 4KB data block, and the fingerprint records for all of the data
blocks in the volume are stored in the fingerprint database file. There is an overhead of less than 2%
associated with this database file.
The size of the deduplication change log files depends on the rate of change of the data and on how
frequently deduplication is run. This accounts for less than 2% overhead in the volume.
Finally, when deduplication is running, it creates some temporary files that could account for up to 2% of
the size of the volume. These temporary metadata files are deleted when the deduplication process has
finished running.
In Data ONTAP 7.2.X, all of the above deduplication metadata files reside in the volume, and this metadata
is therefore captured and locked in the Snapshot copies of the volume as well.
Starting with Data ONTAP 7.3, part of the metadata still resides in the volume, and part of it resides in the
aggregate outside of the volume. The fingerprint database and the change log files are located outside of
the volume in the aggregate and are therefore not captured in Snapshot copies. This change enables
deduplication to achieve higher space savings. However, the other temporary metadata files created during
the deduplication operation are still placed inside the volume. These temporary metadata files are deleted
when the deduplication operation completes. However, if Snapshot copies are created during a
deduplication operation, these temporary metadata files can get locked in Snapshot copies, and they remain
there until the Snapshot copies are deleted.
The guideline for the amount of extra space that should be left in the aggregate or volume for the
deduplication metadata overhead is as follows:
If you’re running Data ONTAP 7.2.X, leave about 6% extra space inside the volume on which you plan
to run deduplication.
If you’re running Data ONTAP 7.3, leave about 2% extra space inside the volume on which you plan to
run deduplication, and around 4% extra space outside the volume in the aggregate, for each volume
running deduplication.
3.4
SPACE SAVINGS ESTIMATION TOOL (SSET)
The actual amount of data space reduction depends on the type of data. For this reason, the SSET should
be used to analyze the actual data set and determine the effectiveness of deduplication on that particular
data set.
When executed, the SSET crawls through all the files in the specified path and estimates the space savings
that will be achieved by deduplication. Although actual deduplication space savings may deviate from what
the estimation tool predicts, use and testing so far indicate that in general, the actual results are within +/ –
5% of the space savings that the tool predicts.
17
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
OVERVIEW OF SSET
The SSET is available to NetApp system engineers, including NetApp partners, and performs nonintrusive
testing of the data set to determine the effectiveness of deduplication.
This tool is intended for use only by NetApp personnel to analyze data at current or prospective NetApp
users. By installing this software, the user agrees to keep this tool and any results from this tool confidential
between them and NetApp.
®
®
The deduplication Space Savings Estimator Tool is available for Linux and Windows systems, which have
the data available locally or using CIFS/NFS. See the SSET readme file for complete usage information.
LIMITATIONS OF THE SSET
The SSET runs on either a Linux system or a Windows system.
It is limited to evaluating 2TB of data or less. If the path given contains more than 2TB, once the tool has
processed the first 2TB of data, the tool indicates that the maximum size has been reached and displays the
results of the data that it has processed until that time (the rest of the data is ignored).
The tool is designed to examine data that is available either locally or using NFS/CIFS only.
For more information about SSET, read the SSET documentation. The SSET tool can be downloaded from
the NetApp internal and PartnerCenter Web sites.
3.5
DEDUPLICATION LIMITATIONS
This section discusses what’s supported and what’s not supported, and the do’s and don’ts. Some of this
information may be covered elsewhere in this report as well.
GENERAL CAVEATS
Deduplication metadata (fingerprint file and change logs) is not deduplicated.
Other metadata (such as directory metadata) is not deduplicated either. Therefore, for heavily replicated
directory environments with a large number of small files (for example, Web space), the amount of space
savings that can be achieved may be low.
Backup of the deduplicated volume using NDMP is supported, but there is no space optimization when the
data is written to tape because it’s a logical operation. (This could actually be considered an advantage,
since in this case the tape does not contain a proprietary format.)
When deduplication is used in an environment where quotas are used, the quotas cannot be oversubscribed
on a volume. For example, a user with a quota limit of 1TB can’t store more than 1TB of data in a
deduplicated volume even if this data fits into less than 1TB of physical space on the storage system.
Storage administrators can use the saved space as desired.
Only data in the active file system is deduplicated. Data pointed to by Snapshot copies that were created
before deduplication is run is not released until the Snapshot copy is deleted.
MAXIMUM FLEXIBLE VOLUME SIZE
The maximum flexible volume size limitation for deduplication varies based on the platform (this number
depends primarily on the amount of system memory). When this limit is reached, writes to the volume fail
just as they would with any other volume after it is full.
This could be important to consider if the flexible volumes are ever moved to a different platform with a
smaller maximum flexible volume size.
Table 5 shows the maximum usable flexible volume size limits (including any snap reserve space) for the
different NetApp storage system platforms. For versions of Data ONTAP prior to 7.3.1, if a volume ever gets
larger than this limit and is later shrunk to a smaller size, deduplication cannot be enabled on that volume.
18
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
Table 5) Maximum deduplicated volume sizes.
Data ONTAP 7.2.X (Starting with 7.2.5.1) and Data ONTAP 7.3.0
FAS2020
FAS3020
FAS3050
FAS3040
N5200
N5500
FAS2050
R200
FAS3070
FAS6030
FAS6070
FAS3140
N5600
FAS6040
FAS6080
N5300
FAS3160
N7600
N7800
FAS3170
0.5TB
1TB
2TB
3TB
4TB
6TB
10TB
16TB
FAS3020
FAS3050
FAS3040
R200
FAS3070
FAS6030
FAS6070
N5200
N5500
FAS3140
N5600
FAS6040
FAS6080
N5300
FAS3160
N7600
N7800
Data ONTAP 7.3.1 or higher
FAS2020
FAS2050
FAS3170
1TB
2TB
3TB
4TB
4TB
16TB
16TB
16TB
The maximum shared data limit per volume for deduplication is 16TB, regardless of the platform type. Once
this limit is reached, there is no more deduplication of data in the volume, but writes to the volume continue
to work successfully until the volume gets completely full.
Table 6 shows the maximum total data limit per deduplicated volume for each platform. This is the maximum
amount of data that can be stored in a deduplicated volume. This limit is equal to the maximum volume size
plus the maximum shared data limit. (For example, in an R200 system that can have a deduplicated volume
of up to 4TB in size, 20TB of data can be stored; that is 4TB + 16TB = 20 TB.)
Table 6) Maximum total data limit in a deduplicated volume.
Data ONTAP 7.2.X (Starting with 7.2.5.1) and Data ONTAP 7.3.0
FAS2020
FAS3020
FAS3050
FAS3040
N5200
N5500
FAS2050
R200
FAS3070
FAS6030
FAS6070
FAS3140
N5600
FAS6040
FAS6080
N5300
FAS3160
N7600
N7800
FAS3170
16.5TB
17TB
18TB
19TB
20TB
22TB
26TB
32TB
FAS3020
FAS3050
FAS3040
R200
FAS3070
FAS6030
FAS6070
N5200
N5500
FAS3140
N5600
FAS6040
FAS6080
N5300
FAS3160
N7600
N7800
Data ONTAP 7.3.1
FAS2020
FAS2050
FAS3170
17TB
19
18TB
19TB
20TB
20TB
32TB
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
32TB
32TB
NUMBER OF DEDUPLICATION PROCESSES
A maximum of eight deduplication processes can be run at the same time on one FAS system.
If another flexible volume is scheduled to have deduplication run while eight deduplication processes are
already running, deduplication for this additional flexible volume is queued. For example, suppose that a
user sets a default schedule (sun-sat@0) for 10 deduplicated volumes. Eight will run at midnight, and the
remaining two will be queued.
As soon as one of the eight current deduplication processes completes, one of the queued ones starts;
when another deduplication process completes, the second queued one starts.
Next time deduplication is scheduled to run on these same 10 flexible volumes, a round-robin paradigm is
used so that the same volumes aren’t always the first ones to run.
With Data ONTAP 7.2.X, for manually triggered deduplication runs, if eight deduplication processes are
already running when a command is issued to start another one, the request fails and the operation is not
queued. However, starting with Data ONTAP 7.3, the manually triggered deduplication runs are also queued
if eight deduplication operations are already running (including the sis start –s command).
4
DEDUPLICATION WITH OTHER NETAPP FEATURES
For the versions of Data ONTAP that are required to run deduplication with the NetApp features described in
this section, read the section on deduplication limitations.
4.1
DEDUPLICATION AND SNAPSHOT COPIES
Deduplication only deduplicates data in the active file system, and that data could be locked in Snapshot
copies created before deduplication, causing reduced storage savings.
There are two types of data that can be locked in Snapshot copies:
Data can be locked in a Snapshot copy if the copy is created before deduplication is run. This effect can
be mitigated by always running deduplication before a Snapshot copy is created.
Deduplication metadata could get locked in a Snapshot copy when the copy is created. In Data ONTAP
7.2.X, all the deduplication metadata resides in the volume. Starting with Data ONTAP 7.3.0, part of the
metadata resides in the volume, and part of it resides in the aggregate outside the volume. The
fingerprint database and the change log files that are used in the deduplication process are located
outside of the volume in the aggregate and are therefore not captured in Snapshot copies. This change
enables deduplication to achieve higher space savings. However, some other temporary metadata files
created during the deduplication operation are still placed inside the volume. These temporary metadata
files are deleted when the deduplication operation completes. (For the size of these temporary
metadata files, see section 3.3.2, ―Deduplication Metadata Overhead.‖) These temporary metadata files
can get locked in Snapshot copies, if the copies are created during a deduplication operation. The
metadata files remain locked until the Snapshot copies are deleted.
For deduplication to provide the most benefit when used in conjunction with Snapshot copies, the following
best practices should be considered:
Run deduplication before creating new Snapshot copies.
Remove unnecessary Snapshot copies maintained in deduplicated volumes.
If possible, reduce the retention time of Snapshot copies maintained in deduplicated volumes.
Schedule deduplication only after significant new data has been written to the volume.
Configure appropriate reserve space for the Snapshot copies.
If the space used by Snapshot copies grows to more than 100%, it will cause df –s to report incorrect
results, because some space from the active file system is being taken away by Snapshot, and
therefore actual savings from deduplication aren’t reported.
If snap reserve is 0, you should turn off the Snapshot auto-create schedule (this is the case in most
LUN deployments).
20
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
4.2
DEDUPLICATION AND SNAPRESTORE
®
The SnapRestore functionality is supported with deduplication, and it works in the same way with
deduplication as it does without deduplication. If you’re running Data ONTAP 7.3, note the following.
Starting with Data ONTAP 7.3, the deduplication metadata files (the fingerprint database and the change log
files) do not get restored when SnapRestore is executed, because they are located outside the volume in
the aggregate. In this case, after the SnapRestore operation, there is not a fingerprint database file in the
active file system for the data. This data, however, retains the original space savings. After SnapRestore, if
deduplication is enabled on the volume, any new data written to the volume continues to be deduplicated.
However, the deduplication process obtains space savings in the new data only and does not deduplicate
between the new data and the restored data. To run deduplication for all the data in the volume (and thus
obtain higher space savings), use the sis start -s command. This command builds the fingerprint
database for all the data in the volume. Depending on the size of the logical data in the volume, this process
can take a long time to complete.
Before using the sis start -s command, make sure that both the volume and the aggregate containing
the volume have sufficient free space to accommodate the addition of the deduplication metadata. For
information about how much extra space to leave for the deduplication metadata, see the section
―Deduplication Metadata Overhead.‖
4.3
DEDUPLICATION AND THE VOL COPY COMMAND
When deduplicated data is copied by using the volume copy command, the copy of the data at the
destination location inherits all the deduplication attributes and storage savings of the original data.
Starting with Data ONTAP 7.3, some of the deduplication metadata files do not get copied by the vol copy
command, because they are located outside of the volume in the aggregate. In this case, there is no
fingerprint database file in the destination volume for the data. However, the data retains the space savings.
The deduplication process also continues for any new data written to the destination volume and creates the
fingerprint database for the new data. The deduplication process obtains space savings in the new data only
and does not deduplicate between the new data and the old data. To run deduplication for all the data in the
cloned volume (and thus obtain higher space savings), use the sis start -s command. Depending on
the size of the logical data in the volume, this process can take a long time to complete.
4.4
DEDUPLICATION AND READ REALLOCATION (REALLOC)
For workloads that perform a mixture of random writes, and large and multiple sequential reads, read
reallocation improves the file layout and the sequential read performance. When you enable read
reallocation, Data ONTAP analyzes the parts of the file that are read sequentially. If the associated blocks
are not already largely contiguous, Data ONTAP updates the file layout by rewriting those blocks to another
location on disk. The rewrite improves the file layout, thus improving the sequential read performance the
next time that section of the file is read. However, read reallocation might result in more storage use if
Snapshot copies are used. It might also result in a higher load on the storage system. If you want to enable
read reallocation but storage space is a concern, you can enable read reallocation on FlexVol® volumes
using the space_optimized option. The space_optimized option conserves space but can slow read
performance through the Snapshot copies. Therefore, if fast read performance through Snapshot copies is a
high priority to you, do not use space_optimized.
A read reallocation scan does not rearrange blocks on disk that are shared between files by deduplication
on deduplicated volumes. Since read reallocation does not predictably improve the file layout and the
sequential read performance when used on deduplicated volumes, performing read reallocation on
deduplicated volumes is not supported. Instead, for files to benefit from read reallocation, they should be
stored on volumes that are not enabled for deduplication.
4.5
DEDUPLICATION AND FLEXCLONE VOLUMES
When a FlexClone volume (cloned volume) is created:
The FlexClone volume of a deduplicated volume is a deduplicated volume.
21
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
The cloned volume inherits the deduplication configuration of the parent volume, such as the
deduplication schedule.
Starting with Data ONTAP 7.3, the deduplication metadata files (the fingerprint database and the
change log files) do not get cloned, because they are located outside the volume in the aggregate. In
this case, there is no fingerprint database file in the cloned volume for the data that came from the
parent. However, the data in the cloned volume inherits the space savings of the original data. The
deduplication process also continues for any new data written to the clone and creates the fingerprint
database for the new data. However, the deduplication process obtains space savings in the new data
only and does not deduplicate between the new data and the old data. To run deduplication for all the
data in the cloned volume (and thus obtain higher space savings), use the sis start -s command.
Depending on the size of the logical data in the volume, this process can take a long time to complete.
Beginning with 7.3.1, in addition to standard FlexClone, FlexClone at the file and LUN level is available
and is allowed on deduplicated volumes.
Deduplication can be used to regain capacity savings on data that was copied using FlexClone at the
file or LUN level, and has been logically migrated (that is, with qtree SnapMirror, SnapVault,
NDMPdump, and so on).
For additional information regarding FlexClone, refer to TR-3742, ―Using FlexClone to Clone Files and
LUNs.‖
VOLUME SPLITTING
When a cloned volume is split from the parent volume, all of the data in the clone that was a part of the
parent volume (that is, not including the data that was written to the clone volume after the clone was
created) gets undeduplicated after the volume split operation. If deduplication is running on the clone
volume, this data gets deduplicated again in subsequent deduplication operations on the volume.
4.6
DEDUPLICATION AND ACTIVE-ACTIVE CONFIGURATION
NetApp active-active controller configurations are supported with deduplication in the following manner:
In active-active state, where both nodes of the cluster are up and no takeover has been done, dedupe
on each node will work independently of one another.
The total number of concurrent deduplication operations allowed on each node of an active-active
configuration is eight.
Writes to the flexible volume have fingerprints written to the change log.
Upon failover to the partner node, there will not be any deduplication process running for the failed
node. However, change logging for the failed node will continue to happen, and upon failback, normal
deduplication operations will resume using the updated change log. The resumed deduplication
processes will start at the times scheduled for each volume or can be started manually.
While in failover mode, the deduplication operations of the local node will not be impacted.
Data ONTAP 7.2.X has no sis administration operations or deduplication function. However, starting
with Data ONTAP 7.3, the following commands are supported for partner volumes in takeover mode:
sis status, sis stat, sis on, sis off.
Also, starting with Data ONTAP version 7.3, for SnapVault with NetBackup™, block sharing is supported
for partner volumes in takeover mode.
Since deduplication is a licensed option behind the NearStore option license, NetApp recommends having
both nodes in an active-active controller configuration licensed with the NearStore option and with
deduplication.
Deduplication does not add any overhead in an active-active configuration other than additional disk I/O.
For additional information regarding active-active configurations, refer to TR-3450, Active/Active Controller
Configuration Overview and Best Practice Guidelines.
4.7
DEDUPLICATION AND V-SERIES
In Data ONTAP 7.3 and later, you can use deduplication on a V-Series system with a NearStore license.
When using NetApp deduplication for FAS with V-Series, the following should be taken into consideration:
22
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
The checksum type you can use with deduplication on a V-Series system is restricted. Only block checksum
type (BCS) is supported with deduplication on a V-Series system. Zoned checksums are not supported, and
a performance degradation will exist on random workloads with zoned checksums (ZCS). Refer to the Data
ONTAP Data Protection Online Backup and Recovery Guide for information about NearStore configuration.
For additional information regarding V-Series systems, refer to TR-3461, Best Practices Guide for V-Series.
4.8
DEDUPLICATION AND SNAPMIRROR REPLICATION
Although there are substantial benefits to be achieved with deduplication alone, a complete storage solution
typically involves the need to additionally mirror the data to another location for disaster recovery purposes.
Replication of the deduplicated volume is supported by using SnapMirror in two ways—volume SnapMirror
and –qtree SnapMirror, as discussed in the next two subsections. Keep in mind that deduplication is
supported only on NetApp storage systems that are running the NearStore option. So any flexible volume
shown in the following figures with deduplication running, even if it’s a SnapMirror ―primary,‖ is on a
NearStore option licensed system.
NetApp recommends not using deduplication with sync SnapMirror. Although technically it will work, the
integration and scheduling of deduplication with sync SnapMirror are complicated to implement in the type of
rigorous real-world scenarios that demand synchronous replication.
For a complete discussion of SnapMirror, refer to TR-3446, SnapMirror Async Best Practices Guide.
REPLICATING WITH VOLUME SNAPMIRROR
A deduplicated volume can be replicated to a secondary storage system (destination) by using volume
SnapMirror, as shown in Figure 3.
Figure 3) Volume SnapMirror replication of a deduplicated volume for disaster recovery.
To run deduplication with volume SnapMirror:
Deduplication must be licensed at the primary location (source). However, the NearStore option must
be licensed on both the source and destination (or an R200 must be used in one or both locations).
Starting with Data ONTAP 7.3.1, the NearStore license is no longer required on the destination system,
but installation of the deduplication license on the destination system is still recommended as part of the
best practices.
Deduplication does not need to be licensed at the destination. However, if the primary site is down and
the secondary location becomes the new primary, deduplication needs to be licensed for continued
deduplication to occur. Therefore the best practice is to have deduplication licensed at both locations.
In a volume SnapMirror relationship, the destination storage system should use an identical or later
release of Data ONTAP.
23
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
Deduplication can be enabled, run, and managed only from the primary location. However, the flexible
volume at the secondary location inherits all the deduplication attributes and storage savings using
SnapMirror.
Shared blocks are transferred only once, so deduplication reduces network bandwidth usage too.
The volume SnapMirror update schedule is not tied to the deduplication schedule.
The maximum volume size limit is imposed based on the lower maximum volume size limit of the
source and destination volumes.
When configuring volume SnapMirror and deduplication, it is important to consider the deduplication
schedule and the volume SnapMirror schedule. As a best practice, start volume SnapMirror transfers of a
deduplicated volume after deduplication has completed (that is, not in the middle of the deduplication
process). This is to avoid sending undeduplicated data and additional temporary metadata files over the
network. If the temporary metadata files in the source volume are locked in Snapshot copies, they also
consume extra space in the source and destination volumes.
Volume SnapMirror performance degradation can increase with deduplicated volumes. This extra overhead
needs to be accounted for when sizing the storage solution. For more information, see the section
―Deduplication Performance.‖
The Impact of Moving Deduplication Metadata Files Outside the Volume
Starting with Data ONTAP 7.3, most of the deduplication metadata resides in the aggregate outside the
volume. Therefore it does not get captured in Snapshot copies, and volume SnapMirror does not replicate
this data. This provides additional network bandwidth savings. However, some temporary metadata files are
still kept inside the volume and are deleted when the deduplication operation completes. If Snapshot copies
are created during the deduplication operation, these temporary metadata files are locked in Snapshot
copies, so a volume SnapMirror update that is initiated during a deduplication process transfers these
temporary metadata files over the network. To prevent this extra data from being replicated, schedule the
volume SnapMirror updates to take place after the deduplication operation has finished running on the
source volume.
In case of a disaster at the primary location, you may need to break the volume SnapMirror relationship and
have the volume SnapMirror destination start serving data. In this case, there is no fingerprint database file
at the destination for the existing data on the destination volume. However, the existing data retains the
space savings from the deduplication operations performed earlier on the original volume SnapMirror
source. Also, the deduplication process continues for new data being written to the volume and creates the
fingerprint database for this new data. The deduplication process obtains space savings in the new data only
and doesn’t deduplicate between the new data and the old data. To run deduplication for all the data in the
volume (and thus obtain higher space savings), use the sis start -s command. This command builds
the fingerprint database for all the data in the volume. Depending on the size of the logical data in the
volume, this process may take a long time to complete.
Important: Before using the sis start -s command, make sure that both the volume and the aggregate
containing the volume have sufficient free space to accommodate the addition of the deduplication
metadata. For information about how much extra space to leave for the deduplication metadata, see section
―Deduplication Metadata Overhead.‖
REPLICATING WITH QTREE SNAPMIRROR
When using qtree SnapMirror with deduplication, remember the following points:
Deduplication can be enabled on the source system, the destination system, or both.
Both the deduplication license and the SnapMirror license must be installed on the system where
deduplication is required.
Unlike volume SnapMirror, no network bandwidth savings are obtained with qtree SnapMirror, because
the source system sends undeduplicated data to the destination system, even if deduplication is
enabled on the source system.
The deduplication schedule is not tied to qtree SnapMirror updates on either the source or the
destination. However, a deduplication schedule can be set up independently of the qtree SnapMirror
schedule. For example, on the destination, the deduplication process does not automatically start at the
completion of qtree SnapMirror transfers.
24
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
As a best practice, NetApp recommends performing qtree SnapMirror updates after the deduplication
process on the source volume has finished running. If a qtree SnapMirror update occurs while the
deduplication process is running on the source volume, then in addition to the transfer of the changed data
blocks, some unchanged data blocks might also get transferred to the destination. If deduplication is not
running on the destination volume, then the redundant data that is transferred occupies extra storage space
on the destination volume.
NetApp also recommends that if deduplication is used on the source volume, then it should also be used on
the destination volume. However, you don’t have to use deduplication on the source volume if you are
planning to use deduplication only on the destination volume.
As far as the qtree SnapMirror base Snapshot copy is concerned, there are typically only a couple of
Snapshot copies on the destination storage system. If Snapshot copies are not retained long term, they are
constantly rotated out and the deduplicated blocks are freed as the Snapshot copies roll off.
If users want to keep Snapshot copies long term (as a replacement for SnapVault, or for other reasons such
as the ability to have writable, reverse, or resync copies in the event of a disaster), it is possible that
deduplicated data can be locked in Snapshot copies for longer periods of time, which reduces the
deduplication storage savings. This situation can arise when users create Snapshot copies manually or by
using snap sched.
The best practice when using qtree SnapMirror with deduplication is to let qtree SnapMirror use the
minimum number of Snapshot copies it requires (essentially, keep the latest version).
Qtree SnapMirror Replication with Deduplication Enabled on the Source Only
A source deduplicated flexible volume can be replicated to a nondeduplicated volume on the destination by
using qtree SnapMirror, as shown in Figure 4.
Figure 4) Qtree SnapMirror replication from a deduplicated source volume to a nondeduplicated destination
volume.
Keep the following points in mind:
Deduplication is licensed only on the source system.
Deduplication is enabled, run, and managed on a flexible volume at the source.
Deduplication doesn’t yield any network bandwidth savings because qtree SnapMirror works at the
logical layer, and it sends undeduplicated data over the network.
The deduplication schedule is not integrated with the qtree SnapMirror update, and vice versa; it must
be configured independently. The completion of a deduplication process doesn’t automatically start a
qtree SnapMirror transfer, and qtree SnapMirror updates don’t trigger the deduplication operation.
Deduplication storage savings are achieved only on the source system.
25
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
Qtree SnapMirror Replication with Deduplication Enabled on the Destination Only
A nondeduplicated flexible volume on the source can be replicated to a deduplicated volume on the
destination by using qtree SnapMirror, as shown in Figure 5.
Figure 5) Qtree SnapMirror replication from a nondeduplicated source volume to a deduplicated destination volume.
Keep the following points in mind:
Deduplication is licensed only on the destination system.
Deduplication is enabled, run, and managed on a flexible volume at the destination.
Deduplication doesn’t yield any network bandwidth savings.
The deduplication schedule is not integrated with the qtree SnapMirror update, and vice versa; it must
be configured independently. The completion of a qtree SnapMirror update doesn’t automatically start a
deduplication operation on the destination, and the deduplication operation doesn’t trigger a qtree
SnapMirror update.
Deduplication storage savings are achieved only on the destination system.
Qtree SnapMirror with Deduplication Enabled on Both the Source and the Destination
A deduplicated flexible volume on the source can be replicated to a deduplicated volume on the destination
by using qtree SnapMirror, as shown in Figure 6.
Figure 6) Qtree SnapMirror replication from a deduplicated source volume to a deduplicated destination volume.
26
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
Keep the following points in mind:
Deduplication is licensed on both the source and the destination.
Deduplication is enabled, run, and managed independently on the source and the destination.
Deduplication doesn’t yield any network bandwidth savings because qtree SnapMirror works at the
logical layer, and it sends undeduplicated data over the network.
Storage savings at the destination are not achieved automatically when qtree SnapMirror updates
(unlike volume SnapMirror), because the data that is sent over the network to the destination is not
deduplicated. This data must be deduplicated again on the destination after qtree SnapMirror has
transferred the data from the source.
The deduplication schedule is not integrated with the qtree SnapMirror update on either the source or
the destination; it must be configured independently.
Storage savings are achieved on both the source and the destination.
4.9
DEDUPLICATION AND SNAPVAULT
The behavior of deduplication with SnapVault is similar to the behavior of deduplication with qtree
SnapMirror, except for the following points. (For information about other aspects of running deduplication
with SnapMirror, see section ―Replicating with SnapMirror.‖)
The deduplication schedule is tied to the SnapVault schedule on the destination system. The
deduplication schedule on the source is not tied to the SnapVault update schedule, and it can be
configured independently on a volume, just like qtree SnapMirror.
Every SnapVault update (baseline or incremental) kicks off the deduplication process on the destination
after the archival Snapshot is created.
The archival Snapshot copy is replaced with a new one after deduplication has finished running on the
destination. (The name of this new Snapshot copy is the same as that of the archival copy, but the
creation time of this copy is changed.)
The deduplication schedule on the destination cannot be configured manually, and the sis start
command is not allowed either. However, the sis start -s command can be run manually on the
destination.
The SnapVault update is not tied to the deduplication operation; that is, a subsequent incremental
update is allowed to run while the deduplication process on the destination volume from the previous
backup is still in progress. In this case, the deduplication process continues to run, but the archival
Snapshot copy does not get replaced after deduplication has finished running.
When using SnapVault, the maximum volume sizes for deduplication for the primary and secondary are
independent of one another. Volumes on each of the systems will need to abide by their respective
maximum volume size limits.
Protection Manager 3.7 and SnapVault integration with deduplication do not work optimally. This is
because deduplication replaces Snapshot copies. As a result, Protection Manager has to wait for
deduplication to finish before renaming Snapshot copies. During the time that Protection Manager waits,
it does not allow clients to list the Snapshot copies or restore from them. This adversely affects the
recovery point objective.
For additional information regarding SnapVault, refer to TR-3487, SnapVault Design and Implementation
Guide.
For additional information regarding Protection Manager, refer to TR-3710, Protection Manager Best
Practice Guide.
4.10 DEDUPLICATION AND SNAPVAULT FOR NETBACKUP
Deduplication is not supported with SnapVault for NetBackup. This applies to both structured (database) and
unstructured (file) data types. The following notes are provided as related information:
Deduplication in a configuration not based on SnapVault for NetBackup (for example, NetBackup
shared or flexible disk option) remains supported.
27
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
4.11 DEDUPLICATION AND MULTISTORE (VFILER)
Starting with Data ONTAP 7.3, deduplication is supported with MultiStore. In Data ONTAP 7.3, the
deduplication commands are available only in the CLI of vFiler0; however, they allow any volume to be
included in the command arguments, regardless of which vFiler unit the volume is associated with.
Beginning with Data ONTAP 7.3.1, the deduplication commands are available in the CLI of each vFiler unit,
allowing each vFiler unit to be configured from within itself.
4.12 DEDUPLICATION AND SNAPLOCK
Starting with Data ONTAP 7.3.1, deduplication is fully supported with SnapLock, including both enterprise
and compliance modes. If implementing SnapLock and NetApp deduplication for FAS, the following items
should be taken into consideration:
A SnapLock volume with files committed to WORM can be deduplicated. Capacity savings will be
similar to savings where the files were not committed to WORM. Both deduplication and subsequent
undeduplication do not result in any changes to the SnapLock attributes or WORM behavior of the
volume or the file.
Deduplication is applied across WORM, WORM append, and non-WORM (normal) files.
Volume restore from a Snapshot copy is only permitted on SnapLock enterprise volumes. When a
volume restore occurs from a Snapshot copy with deduplicated data, the file system returns to the state
at which the Snapshot copy was created, including the state of deduplication, and the (WORM) status of
the volume and the files.
File folding will continue to function, irrespective of the WORM and deduplication status of the files.
For LockVault™: A Snapshot copy is permanent, meaning it can only be deleted after retention period.
There is no archive Snapshot copy created on the secondary until deduplication completes. If
deduplication is still running when next transfer attempts to begin, then the next transfer is delayed. So
deduplication on LockVault can result in the disruption of the transfer schedule on the primary. Avoiding
the mismatched schedule will allow for optimal capacity savings to be recognized.
Autocommit functions irrespective of the deduplication status of the files.
When using qtree SnapMirror, deduplication needs to be run separately on the source and destination.
The WORM property is carried forward by qtree SnapMirror. Switching on WORM or deduplication on
either end has no effect on the qtree SnapMirror transfers. Undoing deduplication will also have no
effect when done on either the source or the destination.
When using volume SnapMirror, the WORM property of the files is carried forward by volume
SnapMirror. Deduplication only needs to be run on the primary. Volume SnapMirror allows the
secondary to inherit the deduplication. Undoing deduplication can only be done after breaking the
volume SnapMirror relationship.
To revert to a previous release on a system hosting a volume that is deduplicated and has WORM data
on it, deduplication must first be undone (undeduplicated).
If you are reverting to a previous release that does not support deduplication with SnapLock volumes,
prior to Data ONTAP 7.3.1, you must first run the ―sis undo‖ command. If the ―sis undo‖ command is not
run prior to the revert operation, then an error message will be displayed stating that ―sis undo‖ must be
performed.
For additional details regarding SnapLock, refer to TR-3263, WORM Storage on Magnetic Disks Using
SnapLock Compliance and SnapLock Enterprise.
4.13 DEDUPLICATION AND METROCLUSTER
Both stretch MetroCluster and fabric MetroCluster are now fully supported with deduplication. When using
MetroCluster with deduplication, the following items should be taken into consideration:
Stretch MetroCluster with deduplication is supported on Data ONTAP 7.2.5.1 and 7.3.1 or later.
Fabric MetroCluster with deduplication is supported on Data ONTAP 7.2.5.1 and 7.3.1 or later.
Deduplication will have an impact on CPU resources as a result of extra disk write operations. The
increase is due to writing to two plexes. On most platforms the impact will be less than 10%. This
28
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
impact will be experienced on low-end systems (for example, 30xx) more than high-end systems (for
example, 6xxx).
In takeover mode, writes to partner flexible volumes will be change logged. The deduplication process
will not run on the partner flexible volumes while in takeover mode. Upon giveback, data in the change
logs will be processed, and data will get deduplicated.
In takeover mode, change logging will continue until the change log is full. This can occur if the node
remains in takeover mode for a long period of time, such as a disaster. All data will continue to be
accessible regardless of change log availability.
A node in takeover mode will take over the servicing of I/Os targeted at the partner volumes, as well as
its change logging. As a result, additional system resources will be consumed, which may require that
the system workload be appropriately adjusted.
Only a subset of deduplication commands for the partner volumes are available in takeover mode. For a
list of these commands, see section ―Deduplication and Active-Active Configuration.‖
Deduplication must be licensed on both nodes.
For additional information regarding MetroCluster, refer to TR-3548, MetroCluster Design and
Implementation Guide.
4.14 DEDUPLICATION AND DATAFORT ENCRYPTION
If implementing DataFort encryption and NetApp deduplication for FAS, the following items should be taken
into consideration:
Encryption removes data redundancy. As a result, encrypted data usually yields extremely low amounts
of capacity savings.
Deduplication can be run on encrypted data, but capacity savings are expected to be 0%.
Since encryption can be run at the share level, it is possible to create a flexible volume where only part
of the data on the volume is encrypted. If deduplication is run on such a volume, 0% capacity savings is
expected on the encrypted data, but it will still be possible to deduplicate the rest of the volume
effectively.
4.15 DEDUPLICATION AND LUNS
When using NetApp deduplication in a file-based (NFS/CIFS) environment, deduplication is straightforward
and automatic; as duplicate blocks are freed, they are marked as available, and the NetApp system
recognizes these free blocks and makes them available to the volume.
Deduplication in a block-based (FCP/iSCSI) LUN environment is slightly more complicated. This is because
of the space guarantees and fractional reservations often used by LUNs. With space guarantees, for
instance, a 500GB LUN that is created consumes exactly 500GB of physical disk space. If the data in the
LUN is reduced through deduplication, the LUN still reserves the same physical space capacity of 500GB,
and the space savings are not apparent to the user.
LUN space guarantees and fractional reserves can be configured so that the use by the NetApp system of
the freed blocks changes depending on the configuration. By varying the values of certain parameters, freed
blocks can be returned to the LUN overwrite reserve, the volume free pool, or the aggregate free pool, or a
combination.
This section describes five common examples of LUN configurations and deduplication behavior, as
summarized in Table 7.
29
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
Table 7) LUN configuration examples (as described below).
A (Default)
B
C
D
E
LUN space guarantee value
Yes
Yes
Yes
No
No
Volume fractional reserve
value
100
1-99
0
Any
Any
Volume thin provisioned?
No
No
No
No
Yes
After deduplication and thin
provisioning (if applicable), free
blocks are returned to:
Fractional
overwrite
reserve
Fractional
overwrite
reserve
+
Volume free
pool
Volume
free pool
Volume
free pool
Aggregate
free pool
DEFINITIONS
Fractional overwrite reserve: The space that Data ONTAP guarantees will be available for overwriting
blocks in a LUN when space guarantee = Yes. Behavior of the fractional reserve space parameter with
deduplication is the same as if a Snapshot copy has been created in the volume and blocks are being
overwritten.
Volume free pool: Refers to the free blocks in the parent volume of the LUN. These blocks can be
assigned anywhere in the volume as needed.
Aggregate free pool: Refers to the free blocks in the parent aggregate of the LUN. These blocks can
be assigned anywhere in the aggregate as needed.
LUN CONFIGURATION EXAMPLES
Configuration A: The Default LUN Configuration
The default configuration of a NetApp LUN follows. (Best practice for all NetApp LUNs is to turn controller
Snapshot off, delete all scheduled Snapshot copies, and set snap reserve to 0.)
1.
LUN space reservation value = on
Default = on
2.
Volume fractional reserve Value = 100
Default = 100%
3.
Volume guarantee = volume
Default = volume
4.
Snap reserve = 0%
Default = 20%
5.
Autodelete = off
Default = off
6.
Autosize = off
Default = off
7.
Try_first = volume_grow
Default = volume_grow
Description: When a LUN containing default values is deduplicated, no apparent savings are observed by
the storage administrator because the LUN by default was ―space reserved‖ when it was created and
fractional reserve was set to 100% in the volume. Any blocks freed through deduplication are allocated to
the fractional reserve area. This configuration means that overwrite to the LUN should never fail, even if it is
overwritten entirely.
Pros and cons: The advantage of this configuration is that Snapshot copies consume less space when
blocks in the active file system are no longer being used. As a result, this volume can hold more Snapshot
copies. The disadvantage of this configuration is that free blocks are not returned to either the free volume
pool or the free aggregate pool. Moreover, there is no direct space saving in the active file system—in fact,
this configuration could consume more space in the volume due to new indirect blocks being written, if no
Snapshot copies exist in the volume and the Snapshot schedule is turned off.
30
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
Note: If Snapshot copies are turned off for the volume (or no copy exists in the volume) this is not a
recommended configuration or volume for deduplication.
Configuration B: LUN Configuration for Shared Volume Space Savings
If the user wants to apply the freed blocks to both the fractional overwrite reserve area and the volume free
pool, this can be accomplished with the following configuration:
1.
LUN space reservation value = on
2.
Volume fractional reserve value = any value from 1 – 99
3.
Volume guarantee = volume
4.
Snap reserve = 0%
5.
Autodelete = off
6.
Autosize = off
7.
Try_first = volume_grow
Description: The only difference between this configuration and configuration A is that the amount of space
reserved for overwrite is based on the fractional reserve value set for the volume. As a result, this
configuration splits the free blocks between fractional overwrite reserve and volume free space. For
instance, if the fractional reserve value is set to 25, 25% of the freed blocks go into fractional overwrite
reserve and 75% of the freed blocks are returned to the volume free pool.
Pros and cons: The advantage of this configuration is that overwrite space reserve does not increase for
every block being deduplicated. Freed blocks are split between volume free pool and fractional reserve. The
disadvantage of this configuration is that overwrites to the LUN beyond the fractional reserve capacity may
fail because freed blocks may have been already allocated. Another disadvantage of this configuration is
that freed blocks stay in the parent volume and cannot be provisioned to any other volumes in the
aggregate.
Note: If Snapshot copies are turned off for the volume (or if no Snapshot copy exists in the volume) and
percentage of savings due to deduplication is less than the fractional reserve, then this is not a
recommended configuration or volume for deduplication.
Configuration C: LUN Configuration for Maximum Volume Space Savings
If the user wants to apply the freed blocks to the volume free pool, this can be accomplished with the
following configuration:
1.
LUN space reservation value = on
2.
Volume fractional reserve value = 0
3.
Volume guarantee = volume
4.
Snap reserve = 0%
5.
Autodelete = off
6.
Autosize = off
7.
Try_first = volume_grow
Description: The only difference between this configuration and configuration B is that the value of
fractional reserve is set to zero. As a result, this configuration "forces" all the freed blocks to the volume free
pool and no blocks are set aside for fractional reserve.
Pros and cons: The advantage of this configuration is that all the freed blocks are returned to the volume
free pool. The disadvantage is that the chance of overwrite failure is higher than with configurations A and B
because no freed blocks are assigned to the fractional overwrite area.
31
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
Configuration D: LUN Configuration for Maximum Volume Space Savings
If the user wants to apply the freed blocks to the volume free pool, this can be accomplished with the
following configuration:
1.
LUN space reservation value = off
2.
Volume rractional reserve value = any value from 0–100
3.
Volume guarantee = volume
4.
Snap reserve = 0%
5.
Autodelete = off
6.
Autosize = off
7.
Try_first = volume_grow
Description: The difference between this configuration and configuration C is that the LUN is not space
reserved. With LUN space guarantees off, the value for volume fractional reserve is ignored for all LUNs in
this volume. From a deduplication perspective, there is no difference between this and the previous
configuration, and all freed blocks go to the volume free pool.
Pros and cons: From a deduplication perspective, this configuration has same advantages and
disadvantages as configuration C.
Configuration E: LUN Configuration for Maximum Aggregate Space Savings
In many cases, the user may prefer to reclaim all freed blocks from the volume and return these blocks to
the aggregate free pool. This is accomplished with the following configuration:
LUN space reservation value = off
Volume fractional reserve value = any value from 0–100
Volume guarantee = none
Snap reserve = 0%
Autodelete = on
Autosize = on
Try_first = volume_grow
Description: This configuration "forces" the free blocks out of the volume and into the aggregate free pool,
where the blocks can be reprovisioned for any other volumes in the aggregate.
Pros and cons: The advantage of this configuration is that it provides the highest efficiency in aggregate
space provisioning. It also uses the thin provisioning features of Data ONTAP, volume autosize and
Snapshot autodelete, to help administer the space in the solution.
The disadvantage of this configuration is that it requires the storage administrator to monitor the free space
available in the aggregates. With volume autosize and Snapshot autodelete turned on, the volume grows
first if space is available in the aggregate; if not, then Snapshot copies are deleted.
32
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
5
DEDUPLICATION AND VMWARE
VMware environments deduplicate extremely well. However, while working out the VMDK and data store
layouts, keep the following points in mind:
Operating system VMDKs deduplicate extremely well because the binary files, patches, and drivers are
highly redundant between virtual machines (VMs). Maximum savings can be achieved by keeping these
in the same volume.
Application binary VMDKs deduplicate to varying degrees. Duplicate applications deduplicate very well;
applications from the same vendor commonly have similar libraries installed and deduplicate somewhat
successfully; and applications written by different vendors don't deduplicate at all.
Application data sets when deduplicated have varying levels of space savings and performance impact
based on application and intended use. Careful consideration is needed, just as with nonvirtualized
environments, before deciding to keep the application data in a deduplicated volume.
Transient and temporary data such as VM swap files, pagefiles, and user and system temp directories
do not deduplicate well and potentially add significant performance pressure when deduplicated.
Therefore NetApp recommends keeping this data on a separate VMDK and volume that are not
deduplicated.
Data ONTAP 7.3.1 includes a performance enhancement referred to as warm cache extension for zero
blocks. This is particularly applicable to VM environments, where multiple blocks are set to zero as a
result of system initialization. These zero blocks are all recognized as duplicates and are deduplicated
very efficiently. The warm cache extension enhancement provides increased sequential read
performance for such environments, where there will be very large amounts of deduplicated blocks.
Examples of sequential read applications that will benefit from this performance enhancement include
NDMP, SnapVault, some NFS-based application, and dump. This performance enhancement is also
beneficial to the boot-up processes in VDI environments.
The expectation is that about 30% space savings will be achieved overall. This is a conservative number,
and in some cases users have achieved savings of up to 80%. The major factor that affects this percentage
is the amount of application data. New installations typically deduplicate extremely well, because they do not
contain a significant amount of application data.
Important: In VMware, the need for proper partitioning and alignment of the VMDKs is extremely important
(not just for deduplication). VMware must be configured so that the VMDKs are aligned on WAFL 4K block
boundaries as part of a standard VMware implementation. To help prevent the negative performance impact
of LUN/VMDK misalignment, read TR-3428, ―NetApp and VMware Best Practices Guide,‖ at
http://media.netapp.com/documents/tr-3428.pdf. Also note that the applications in which the performance is
heavily affected by deduplication (when these applications are run without VMware) are likely to suffer the
same performance impact from deduplication when they are run with VMware.
A deduplication and VMware solution on NFS is easy and straightforward. Combining deduplication and
VMware with LUNs requires a bit more work. For more information on this, see section 4.10, ―Deduplication
and LUNs.‖
The following subsections describe the different ways that VMware can be configured. For more information
about NetApp storage in a VMware environment, see TR-3428, NetApp and VMware Virtual Infrastructure 3
Storage Best Practices .
5.1
VMFS DATA STORE ON FIBRE CHANNEL OR ISCSI: SINGLE LUN
This is the default configuration, and it’s the way that a large number of VMware installations are done
today. Deduplication occurs across the numerous VMDKs.
33
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
Figure 7) VMFS data store on Fibre Channel or iSCSI—single LUN.
5.2
VMWARE VIRTUAL DISKS OVER NFS/CIFS
This is a new configuration that became available starting with VMware 3.0. It has a low installed base
currently, but it is hot and growing. It is the easiest to configure and allows deduplication to provide the most
space savings.
Figure 8) VMware virtual disks over NFS/CIFS.
34
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
5.3
DEDUPLICATION ARCHIVE OF VMWARE
Deduplication has proven very useful in VMware archive environments. Figure 9 shows an example.
Figure 9) Archive of VMware with deduplication.
Detailed specifications for the example shown in Figure 9:
In this environment, VMware is done using NFS.
This environment uses approximately 1,800 clone copies of their master VMware image. These images
are used to create virtual machines for primary applications and for test and development purposes.
All 1,800 clone copies (~32TB) are stored on a FAS3070 in the Houston data center.
The data is mirrored to the remote site in Austin for disaster recovery.
Once per hour, the FAS3070 images are transferred to an R200 by using SnapMirror.
Deduplication is run nightly on the R200, and the VMware images are reduced in size by 80 to 90%.
6
DEDUPLICATION AND SHAREPOINT
If SharePoint® and NetApp deduplication for FAS will be used together, the following should be taken into
consideration:
Make sure that there is space available in the volume before using the ―sis on‖ command. If the ―sis on‖
command is used on a flexible volume that already has data and is completely full, it will fail. Up to 6%
of the total data size is needed for deduplication of metadata files.
Deduplication is transparent to SharePoint. The block-level changes are not recognized by SharePoint,
so the SharePoint database will be unchanged in size, even though there are capacity savings at the
volume level.
35
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
7
DEDUPLICATION AND EXCHANGE
If Exchange and NetApp deduplication for FAS will be used together, the following should be taken into
consideration:
In some Exchange environments, extents are enabled to improve performance of database validation.
Enabling extents does not rearrange blocks on disk that are shared between files by deduplication on
deduplicated volumes. Since enabling extents does not predictably optimize sequential data block
layout when used on deduplicated volumes, there is no reason to enable extents on deduplicated
volumes.
For additional details regarding Exchange, refer to TR-3578, Microsoft Exchange Server 2007 Best
Practices Guide.
8
DEDUPLICATION AND TIVOLI STORAGE MANAGER (TSM)
If Tivoli Storage Manager (TSM) and NetApp deduplication for FAS will be used together, the following
should be taken into consideration:
Deduplication savings with TSM will not be optimal due to the fact that TSM does not block align data
when it writes files out to its volumes. The net result is that there are less duplicate blocks available to
deduplicate.
TSM compresses files backed up from clients to preserve bandwidth. Compressed data does not
usually yield good savings when deduplicated.
TSM client-based encryption will result in data with no duplicates. Encrypted data does not usually yield
good savings when deduplicated.
TSM’s progressive backup methodology backs up only new or changed files, which reduces the number
of duplicates, since there are not multiple full backups to consider.
9
DEDUPLICATION AND SYMANTEC BACKUP EXEC
If Symantec™ Backup Exec™ and NetApp deduplication for FAS will be used together, the following should
be taken into consideration:
Deduplication savings with Backup Exec will not be optimal due to the fact that Backup Exec does not
block align data when it writes files out to its volumes. The net result is that there are less duplicate
blocks available to deduplicate.
10 DEDUPLICATION AND LOTUS DOMINO
If Lotus Domino and NetApp deduplication for FAS will be used together, the following should be taken into
consideration:
There have been reports of degradation in read performance when deduplication is used with Lotus
Domino on primary storage. This is a note for caution at this time, while the cause behind this behavior
continues to be researched.
36
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
11 TROUBLESHOOTING
This section covers issues that occasionally come up when configuring and running deduplication.
11.1 LICENSING
Make sure that deduplication is properly licensed and, if the platform is not an R200, make sure that the
NearStore option is also properly licensed:
fas3070-rtp01> license
…
a_sis <license key>
nearstore_option <license key>
…
If licensing is removed or expired, no additional deduplication can occur, and no sis commands can run.
However, the flexible volume remains a deduplicated volume, the existing storage savings are kept, and all
data is usable.
Deduplication must be disabled before removing the deduplication license. Before removing the
deduplication license, you must disable deduplication on all the flexible volumes, using the sis off
command. If you attempt to remove the license without first disabling deduplication, you will receive a
warning message asking you to disable this feature. Note: Any volume deduplication that occurred before
removing the license will remain unchanged.
11.2 VOLUME SIZES
From a deduplicated volume size limit perspective, a volume cannot exceed the size limit for the entire life of
the volume (that is, if a volume is larger than the maximum size and is then shrunk, you still cannot enable
deduplication on that volume). If you need to run deduplication on a volume that was (at some point in time)
larger than the maximum supported size, you can do so by creating a new volume and migrating the data
from the old volume to the newly created volume.
Here is an example of the message displayed if the volume is, or has been, too large to enable
deduplication:
london-fs3> sis on /vol/projects
Volume or maxfiles exceeded max allowed for SIS: /vol/projects
11.3 LOGS AND ERROR MESSAGES
The location of the deduplication log file is:
/etc/log/sis
Error messages with explanations:
Registry errors:
Check if vol0 is full (only in Data ONTAP 7.2.X).
Metafile op errors:
Check if the deduplicated volume is full (in Data ONTAP 7.2.X).
Metafile op errors:
Check if the deduplicated aggregate is full (in Data ONTAP 7.2.X).
License errors:
Check if the license is installed.
Change log full error:
Perform a sis start operation to empty the change log metafile when finished
11.4 NOT SEEING SPACE SAVINGS
If you’ve run deduplication on a flexible volume that you’re confident contains data that should deduplicate
well, but you are not seeing any space savings, there’s a good chance that a bunch of Snapshot copies exist
37
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
and are locking a lot of data. This tends to happen especially when deduplication is run on existing flexible
volumes of data.
Use the snap list command to see what Snapshot copies exist and the snap delete command to
remove them. Alternatively, wait for the Snapshot copies to expire and the space savings to appear (see
section 4.1, ―Deduplication and Snapshot Copies‖).
11.5 UNDEDUPLICATING A FLEXIBLE VOLUME
It is possible, and relatively easy, to undeduplicate a flexible volume that has deduplication enabled by
backing out deduplication and turning it back into a regular (nondeduplicated) flexible volume. This can be
done while the flexible volume is online, as described below.
Turn deduplication off on the flexible volume.
Note: This command stops fingerprints from being written to the change log as new data is written to the
flexible volume. If this command is used, and then deduplication is turned back on for this flexible volume,
the flexible volume must be rescanned with the sis start –s command.
sis off <flexvol>
1
Use the following command to recreate the duplicate blocks in the flexible volume:
sis undo <flexvol>
When this command completes, it deletes the fingerprint file and the change log files.
Here is an example of undeduplicating a flexible volume:
r200-rtp01> df –s /vol/VolReallyBig2
/vol/VolReallyBig2/
20568276
3768732
15%
r200-rtp01> sis status /vol/VolReallyBig2
Path
State
Status
Progress
/vol/VolReallyBig2
Enabled
Idle
Idle for 11:11:13
r200-rtp01> sis off /vol/VolReallyBig2
SIS for "/vol/VolReallyBig2" is disabled.
r200-rtp01> sis status /vol/VolReallyBig2
Path
State
Status
Progress
/vol/VolReallyBig2
Disabled
Idle
Idle for 11:11:34
r200-rtp01> sis undo /vol/VolReallyBig2
Wed Feb 7 11:13:15 EST [wafl.scan.start:info]: Starting SIS volume scan
on volume VolReallyBig2.
r200-rtp01> sis status /vol/VolReallyBig2
Path
State
Status
Progress
/vol/VolReallyBig2
Disabled
Undoing
424 MB Processed
r200-rtp01> sis status /vol/VolReallyBig2
No status entry found.
r200-rtp01> df -s /vol/VolReallyBig2
Filesystem
used
saved
%saved
/vol/VolReallyBig2/
24149560
0
0%
Note: If sis undo starts processing and then there is not enough space to undeduplicate, it stops, sends a
message about insufficient space, and leaves the flexible volume deduplicated. Use df –s to understand how
much free space you really have, and then delete either data or Snapshot copies to provide the needed free
space.
1
The undo option of the sis command is available only in the diag mode, accessed using the priv set
diag command.
38
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
11.6 ADDITIONAL REPORTING WITH SIS STAT -1
For additional status information, you can use priv set diag and then use the sis stat –l command
for long, detailed listings. The following are some additional details around the sis stat command:
When volume-name is omitted, it executes for all known SIS volumes.
–l lists all the details about the volume.
-b shows the disk space usage and saved disk space in number of blocks.
-v shows the disk space usage and saved disk space in number of bytes; if stat command executed
without any option it runs with iv option only.
-g lists all the information about dedup block and change log buffer status.
–lv generates reference histograms which can be used for troubleshooting efforts if instructed by
NetApp Global Support.
Block sharing histogram
The block sharing histogram gives an indication of how many shared blocks are contiguous (adjacent to
one another) in nature.
Refcount histogram
The refcount histogram shows the total number of refcounts. That is, it shows the number of times there
are blocks with one reference pointer to them, then the number of times there are blocks with two
references to them, then the number of blocks with three references to them, and so on.
12 ADDITIONAL READING AND REFERENCES
TR-3446, SnapMirror Async Best Practices Guide
TR-3705, NetApp and VMware VDI Best Practices
TR-3702, NetApp and Microsoft Virtualization Best Practices
TR-3742, Using FlexClone to Clone Files and LUNs
TR-3701, NetApp and Microsoft Virtualization: Making Integrated Server and Storage Virtualization a Reality
TR-3694, NetApp and Citrix XenServer 4.1: Building a Virtual Infrastructure from Server to Storage
TR-3428, NetApp and VMware Virtual Infrastructure 3 Storage Best Practices
WP-7053, The 50% Virtualization Guarantee Program Technical Guide
TR-3465, SnapVault for NetBackup Deployment and Implementation Guide
TR-3487, SnapVault Design and Implementation Guide
TR-3483, Thin Provisioning in a NetApp SAN or IP SAN Enterprise Environment
TR-3548, MetroCluster Design and implementation
TR-3263, WORM Storage on Magnetic Disks Using SnapLock Compliance and SnapLock Enterprise
NetApp Data Online Backup and Recovery Guide
39
NetApp Deduplication for FAS and V-Series Deployment and Implementation Guide
13 ADDITIONAL ASSISTANCE
For additional support contact one of the following:
Your local account team
Systems engineer
Account manager
NetApp Global Services
http://now.netapp.com
888.4.NETAPP (United States and Canada)
00.800.44.NETAPP (EMEA/Europe)
+800.800.80.800 (Asia/Pacific)
www.netapp.com
40
© 2009 NetApp. All rights reserved. Specifications are subject to change without notice. NetApp, the NetApp logo, Go further, faster,
Data ONTAP, FlexClone, FlexShare, FlexVol, LockVault, MultiStore, NearStore, SnapLock, SnapMirror, SnapRestore, Snapshot,
SnapVault, vFiler, and WAFL are trademarks or registered trademarks of NetApp, Inc. in the United States and/or other countries.
Linux is a registered trademark of Linus Torvalds. Microsoft, Windows, and SharePoint are registered trademarks of Microsoft.
Solaris is a trademark
Sun
Microsystems,
Inc. Symantec,
Backup Exec,Guide
and NetBackup are trademarks of Symantec Corporation.
NetApp Deduplication
for FAS of
and
V-Series
Deployment
and Implementation
VMware is a registered trademark of VMware, Inc. All other brands or products are trademarks or registered trademarks of their
respective holders and should be treated as such. TR-3505-0309
Download PDF