Replication Director (NetApp) Best Practices -...

55
NetBackup Customer Forum 2013 1 Replication Director (NetApp) – Best Practices Arun Birdawade Srinivasa Murthy

Transcript of Replication Director (NetApp) Best Practices -...

NetBackup Customer Forum 2013 1

Replication Director (NetApp) – Best Practices

Arun Birdawade Srinivasa Murthy

Roadmap Disclaimer!

• This forward-looking indication of plans for products is preliminary and all future release dates are tentative and are subject to change. Any future

release of the product or planned modifications to product capability, functionality or feature are subject to ongoing evaluation by Symantec, and

may or may not be implemented and should not be considered firm commitments by Symantec, and should not be relied upon in making

purchasing decisions.

2 NetBackup Customer Forum 2013

Agenda

NetBackup Customer Forum 2013 3

Typical Replication Director Deployment 1

Replication Director configuration work flow 2

Best Practices 3

Q & A 4

Typical Replication Director Deployment

NetBackup Customer Forum 2013 4

Production Site Replication Site

Application

server

Netapp

DFM NFS, CIFS ,

FC, ISCI

Snapvault/

Snapmirror

Typical Replication Director Deployment

NetBackup Customer Forum 2013 5

Production Site Replication Site

Netbackup Master

Netbackup

client

Netbackup

media

Netapp

DFM NFS, CIFS ,

FC, ISCI

Snapvault/

Snapmirror

nbuplugin

Policy

Typical Replication Director Deployment

NetBackup Customer Forum 2013 6

Production Site Replication Site

Netbackup Master

Netbackup

client

Netbackup

media

Netapp

DFM NFS, CIFS ,

FC, ISCI

Snapvault/

Snapmirror

nbuplugin

Policy

Snapshot Replica

Typical Replication Director Deployment

NetBackup Customer Forum 2013 7

Production Site Replication Site

Netbackup Master

Netbackup

client

Netbackup

media

Netapp

DFM

nbuplugin

Snapvault/

Snapmirror

Policy

Snapshot Replica

NFS, CIFS ,

FC, ISCI

Disk

Tape

Replication Director Configuration Flow – Quick Overlook

1 • Create Storage Server

2 • Import Relations

3 • Create Disk pool and Storage Unit

4 • Configure Storage life cycle policy

5 • Configure Backup Policy

NetBackup Customer Forum 2013 8

Create Storage Server for NetApp DFM

NetBackup Customer Forum 2013 9

nbuplugin

Snapvault/

Snapmirror

Netapp

DFM

NetBackup Master

1. Register NetApp DFM as openstorage server

2. Assign media server for communication with nbuplugin

NetBackup Media

OSTplugin

Import Relationships

NetBackup Customer Forum 2013 10

nbuplugin

Snapvault/

Snapmirror

Netapp

DFM

Netbackup Master

Import tool

1. Generate XML report

2. Generate NetBackup script

SLP Policy STU

3. Create netbackup objects and policies

Create Disk pool and Storage Unit

NetBackup Customer Forum 2013 11

Aggregate1 Aggregate2

Resource Pool 1

Filer1 Filer2

Resource Pool 2

NetBackup Master NetApp DFM

LSU – Resource pool1

Disk Pool 1

Storage Unit 1

LSU – Resource pool2

Disk Pool 2

Storage Unit 2

Configure Storage life cycle

NetBackup Customer Forum 2013 12

1 week 1 month 1 year 1 day

1. Based on retention type, Replication method is decided

2. Hierarchal combination of operations in SLP decide replication topology

Storage life cycle for Mirror Vault cascade topology

NetBackup Customer Forum 2013 13

Primary Secondary SnapMirror

Tertiary SnapVault

Storage life cycle for Mirror Vault fanout topology

NetBackup Customer Forum 2013 14

Primary

Secondary

Secondary

Configure Backup Policy

• Define policy Type • Standard

• MS-Windows

• NDMP

• Standard/MS-Windows with NDMP data mover

• Define frequency of taking snapshots and replication • Frequency or Calendar based schedule

• Incremental, full, synthetic

• Provide Client list to be protected • NAS Filer in case of NDMP policy type

• Unix or Windows Clients for other policy types

• Provide Backup selections • NFS mount for Unix clients

• UNC path for Windows Clients

• Volume/Qtree of NAS Filer in case of NDMP policy.

NetBackup Customer Forum 2013 15

Best Practices

NetBackup Customer Forum 2013 16

Best practices

• Configuring Replication

• Creating Storage Server

• Creating Disk Pools and Storage Units

• Use of Storage unit groups for protecting data spanned across DFM domains

• Managing Replication

• Using SLP Window for replication

• Different RPO requirements for snapshot and Replication

• Throttling Parallel Replication

• Backups from Snapshot and Replication

• Use of alternate client for off-host backup and indexing

• Using accelerator enabled policy

• Using multi-streaming

NetBackup Customer Forum 2013 17

Best practices – Assigning Media servers

• Create Storage Server

– Single media server per DFM domain

– Two Media servers for load balancing

– Single media server for multiple DFM domains

NetBackup Customer Forum 2013 18

NetBackup Master Media Server

nbuplugin

Netapp

DFM

Best practices – Assigning Media servers

• Create Storage Server

– Single media server per DFM domain

– Two Media servers for load balancing,

– Single media server for multiple DFM domains

NetBackup Customer Forum 2013 19

NetBackup Master

Media Server

nbuplugin

Netapp

DFM

Not recommended as Media server doesn’t carry any load in Replication Director environment

Best practices – Assigning Media servers

• Create Storage Server

– Single media server per DFM domain

– Two Media servers for load balancing

– Single media server for multiple DFM domains

NetBackup Customer Forum 2013 20

NetBackup Master Media Server

nbuplugin

Netapp

DFM

nbuplugin

Netapp

DFM

Best practices – Creating Disk pool and Storage Unit

• Multiple LSU’s in a disk pool

• Single LSU in disk pool

• Storage Unit groups

NetBackup Customer Forum 2013 21

Aggregate1

Resource Pool 1

NetBackup Master NetApp DFM

LSU – Resource pool1

Disk Pool 1

Storage Unit 1

Aggregate2

Resource Pool 2 LSU – Resource pool2

Not recommended as additional LSU’s would never get used

Best practices – Creating Disk pool and Storage Unit

• Multiple LSU’s in a disk pool

• Single LSU in disk pool

• Storage Unit groups

NetBackup Customer Forum 2013 22

NetBackup Master NetApp DFM

Aggregate1 Aggregate2

Resource Pool 1

LSU – Resource pool1

Disk Pool 1

Storage Unit 1

DFM has mechanism to use additional storage under resource pool if primary storage gets full

Protecting data spanned across DFM domains – Case1.

NetBackup Customer Forum 2013 23

Application

server1

Netapp

DFM1

Application

server2

Netapp

DFM2

Single NetBackup Domain

• Data is spanned across DFM domains for NetBackup Domain but Application Servers using Data from single DFM domain

•Follow normal configuration procedure and create storage lifecycle policy and backup policy separately for each client.

Protecting data spanned across DFM domains – Case2

NetBackup Customer Forum 2013 24

Application

server1

Netapp

DFM1 Netapp

DFM2

Source Filer1 Source Filer2

Target Filer2 Target Filer1

• Data is spanned across DFM domains for NetBackup Domain as well as Application server

Challenge – How to communicate with both DFM servers simultaneously to create consistent snapshot/replica copy.

Solution – Use STU groups

•Create “Source STU Group” for source filer1 and filer2.

•Create “Target STU Group” for Target filer1 and filer2.

•Use “Source STU Group” in Storage life cycle policy for snapshot operation

•Use “Target STU Group” in Storage life cycle policy for Replication operation

Note – Common Media server need to be configured for both DFM domains

Best Practices – Managing Replication

NetBackup Customer Forum 2013 25

Using SLP window for managing replication

NetBackup Customer Forum 2013 26

Using SLP window for managing Replication

NetBackup Customer Forum 2013 27

Application

server

Source Volume

S1 – 8AM

S2 – 9AM

S3 – 10AM

S4 – 11AM

S12 – 9PM

Nightly Replication

Target Volume

S1 – 8AM

S2 – 9AM

S3 – 10AM

S4 – 11AM

S12 – 9PM

Production Site Replication Site

Different RPO requirements for snapshot and Replication

NetBackup Customer Forum 2013 28

Different RPO requirements for snapshot and Replication

NetBackup Customer Forum 2013 29

Different RPO requirements for snapshot and Replication

NetBackup Customer Forum 2013 30

NetBackup Customer Forum 2013 31

Application

server

Source Volume

S1 – 8AM

S2 – 9AM

S3 – 10AM

S4 – 11AM

S12 – 9PM

Replication

Target Volume

S12 – 9PM

Production Site Replication Site

Different RPO requirements for snapshot and Replication

Throttling parallel replications

NetBackup Customer Forum 2013 32

Throttling parallel replications

NetBackup Customer Forum 2013 33

NetBackup Job manager

1. Replication for STU1

2. Replication for STU1

3. Replication for STU2

4. Replication for STU2

Data Fabric Manager

Concurrent Jobs = 2

Max Concurrent jobs STU1

Max Concurrent jobs STU2

1. Replication for STU1

2. Replication for STU1

3. Replication for STU2

4. Replication for STU2

STU1 – Target Filer1

STU2 – Target Filer2

1. Replication for STU1

2. Replication for STU1

3. Replication for STU2

4. Replication for STU2

Throttling parallel replications

NetBackup Customer Forum 2013 34

NetBackup Job manager

1. Replication for STU1

2. Replication for STU1

3. Replication for STU2

4. Replication for STU2

Data Fabric Manager

Concurrent Jobs = 2

Max Concurrent jobs STU1

Max Concurrent jobs STU2

1. Replication for STU1

2. Replication for STU1

3. Replication for STU2

4. Replication for STU2

STU1 – Target Filer1

STU2 – Target Filer2

1. Replication for STU1

2. Replication for STU1

3. Replication for STU2

4. Replication for STU2

* Replication Load only on FIler1

Throttling parallel replications

NetBackup Customer Forum 2013 35

NetBackup Job manager

1. Replication for STU1

2. Replication for STU1

3. Replication for STU2

4. Replication for STU2

Data Fabric Manager

Concurrent Jobs = 2

Max Concurrent jobs STU1

Max Concurrent jobs STU2

1. Replication for STU1

2. Replication for STU1

3. Replication for STU2

4. Replication for STU2

STU1 – Target Filer1

STU2 – Target Filer2

1. Replication for STU1

2. Replication for STU1

3. Replication for STU2

4. Replication for STU2

* Replication Load only on FIler2

Throttling parallel replications

NetBackup Customer Forum 2013 36

NetBackup Job manager

1. Replication for STU1

2. Replication for STU1

3. Replication for STU2

4. Replication for STU2

Data Fabric Manager

Concurrent Jobs = 2

Max Concurrent jobs STU1=1

Max Concurrent jobs STU2=1

1. Replication for STU1

3. Replication for STU2

STU1 – Target Filer1

STU2 – Target Filer2

1. Replication for STU1

3. Replication for STU2

Throttling parallel replications

NetBackup Customer Forum 2013 37

NetBackup Job manager

1. Replication for STU1

2. Replication for STU1

3. Replication for STU2

4. Replication for STU2

Data Fabric Manager

Concurrent Jobs = 2

Max Concurrent jobs STU1=1

Max Concurrent jobs STU2=1

1. Replication for STU1

3. Replication for STU2

STU1 – Target Filer1

STU2 – Target Filer2

1. Replication for STU1

3. Replication for STU2 * Replication Load distributed across filers

Throttling parallel replications

NetBackup Customer Forum 2013 38

NetBackup Job manager

2. Replication for STU1

4. Replication for STU2

Data Fabric Manager

Concurrent Jobs = 2

Max Concurrent jobs STU1=1

Max Concurrent jobs STU2=1

2. Replication for STU1

4. Replication for STU2

STU1 – Target Filer1

STU2 – Target Filer2

2. Replication for STU1

4. Replication for STU2

Throttling parallel replications

NetBackup Customer Forum 2013 39

NetBackup Job manager

2. Replication for STU1

4. Replication for STU2

Data Fabric Manager

Concurrent Jobs = 2

Max Concurrent jobs STU1=1

Max Concurrent jobs STU2=1

2. Replication for STU1

4. Replication for STU2

STU1 – Target Filer1

STU2 – Target Filer2

2. Replication for STU1

4. Replication for STU2

* Replication Load distributed across filers

Throttling parallel replications

NetBackup Customer Forum 2013 40

Throttling parallel replications

NetBackup Customer Forum 2013 41

NetBackup Job manager

1. Replication for STU1

2. Replication for STU1

3. Replication for STU2

4. Replication for STU2

Data Fabric Manager

Concurrent Jobs = 2

Max Concurrent jobs STU1=1

Max Concurrent jobs STU2=1

STU1 – Target Filer1

STU2 – Target Filer2

(Image1, Image2)

1. Replication for STU1

3. Replication for STU2

(Image1, Image2)

1. Image1 Replication STU1

3. Replication for STU2

2. Image2 Replication STU1

Throttling parallel replications

NetBackup Customer Forum 2013 42

NetBackup Job manager

1. Replication for STU1

2. Replication for STU1

3. Replication for STU2

4. Replication for STU2

Data Fabric Manager

Concurrent Jobs = 2

Max Concurrent jobs STU1=1

Max Concurrent jobs STU2=1

STU1 – Target Filer1

STU2 – Target Filer2

(Image1, Image2)

1. Replication for STU1

3. Replication for STU2

(Image1, Image2)

1. Image1 Replication STU1

3. Replication for STU2

2. Image2 Replication STU1 * Replication Load only on FIler1

Throttling parallel replications

NetBackup Customer Forum 2013 43

NetBackup Job manager

1. Replication for STU1

2. Replication for STU1

3. Replication for STU2

4. Replication for STU2

Data Fabric Manager

Concurrent Jobs = 2 STU1 – Target Filer1

STU2 – Target Filer2

(Image1, Image2)

1. Replication for STU1

3. Replication for STU2

(Image1)

1. Image1 Replication STU1

3. Replication for STU2

Max Concurrent jobs STU1=1

Max Concurrent jobs STU2=1

SLP : Max images per replication = 1

* Replication Load distributed across filers

Best Practices – Backup/Index from Replication

NetBackup Customer Forum 2013 44

Effective use of Indexing Snapshot

• Indexing updates the catalog file (.f file) which allows the users to search for individual files for restore

Advantages : Snapshots wouldn’t be mounted while searching for individual files

How :

• Configured in the SLP Policy. Can be configured after Primary Snapshot or a Replica Copy.

Best Practices:

• Index from replica snapshot whenever possible to reduce load on production filers.

• Index from snapshot/replica copy if the backup to tape/disk is not scheduled immediately after replication

• Use Alternate client to reduce load on production servers.

NetBackup Customer Forum 2013 45

Use of Alternate client for backup and index

NetBackup Customer Forum 2013 46

Using multi-streaming for tape backups

NetBackup Customer Forum 2013 47

• Divides single large backup in to multiple concurrent data streams.

•Reduces time window for duplication to tape from snapshot/Replica copy

•Reduces retry time in the event of duplication failures

Using Accelerator along with Replication Director

NetBackup Customer Forum 2013 48

• Useful in case duplication is planned from snapshot/replica to Netbackup Media Server Deduplication Pool.

•Track logs are used to intelligently identify changed files and changed segments in to files.

•Provides Full backups at the cost of Incremental Backups

Team Learnings

NetBackup Customer Forum 2013 49

NetBackup Customer Forum 2013 50

Flexclone – When are they created?

When all of the following conditions are met:

• Import of Replica copy

• Either it is vault copy or mirror copy following vault(ie PVM topology)

• Policy having one or more volumes as backup selection which contain atleast one qtree

Import of Snapshot – When?

1. Browse

2. Restore

3. Backup from Snapshot

4. Indexing

Import of Snapshot takes longer time due to Flexclone creation

Which Replication topology does it affect ?

NetBackup Customer Forum 2013 51

Primary Secondary SnapMirror

Primary Secondary SnapVault

Primary Secondary SnapMirror

Tertiary SnapMirror

Primary Secondary SnapMirror

Tertiary SnapVault

No Impact on Import performance

Impact on Import performance

Which Replication topology does it affect ?

NetBackup Customer Forum 2013 52

Primary Secondary SnapVault

Tertiary SnapMirror

Primary

Secondary

Secondary

No Impact on Import performance

Impact on Import performance

NetBackup Customer Forum 2013 53

Possible alternatives • Pros and Cons

– Qtree centric backup selection

• Mount the qtrees individually on the host

• + speeds up import of copy2 snapshots for indexing, browse, restore

• + Can use wildcards in the backup selection with qtrees mounted inside a directory

• - requires mounts of qtress on a client (if not using NDMP policy type)

• = Limitations on the number of qtree (& the mount points) that can exist in the policy.

• - Cannot protect data residing on the volume directly (not inside any qtree).

– Index off of primary snapshot copy

• + creates .f file, which speeds browse for all copies.

• - index operation puts load on production filers.

• - does not speed up imports needed at restore time.

– NDMP policy type.

• + Will save the effort of mounting the volumes/qtrees on the host.

• + Browse will not result in mounting the snapshot/replica on the host.

– Still the flexclone will be created (wherever applicable)

– SnapMirror instead of SnapVault for copy2

• + easier to manage. replication by volume, volumes instead of qtrees in policies

• - may require more filer resources

Performance Numbers

• Each qtree contains 1 Million files

NetBackup Customer Forum 2013 54

Thank you!

Copyright © 2013 Symantec Corporation. All rights reserved. Symantec and the Symantec Logo are trademarks or registered trademarks of Symantec Corporation or its affiliates in the U.S. and other countries. Other names may be trademarks of their respective owners. This document is provided for informational purposes only and is not intended as advertising. All warranties relating to the information in this document, either express or implied, are disclaimed to the maximum extent allowed by law. The information in this document is subject to change without notice.

NetBackup Customer Forum 2013 56

Arun Birdawade, Srinivasa Murthy