Chapter 01 - NetWorker Overview
Chapter 01 - NetWorker Overview
Chapter 01 - NetWorker Overview
In this chapter you learn about NetWorker components including host roles, control data,
daemons (UNIX) or processes started by NetWorker services (Windows), and the role of each
component in providing backup protection for your important data. You also learn about the
advantages of using EMC NetWorker®. Lastly, a brief overview of the backup process will be
presented.
NetWorker Hosts
Processes and Functions
App
App Oracle
SQL
SQL App
EMC
IT App
App NetWorker
Solution
Operating
Operating
System
System
Hardware
Hardware
There are many pieces in an overall solution for your business computing needs. EMC
NetWorker works within the existing framework of hardware, operating system software, and
network communication protocols to provide protection for the critical data that the framework
supports.
NetWorker protects critical business data by centralizing, automating and accelerating backup
and recovery operations across an enterprise. NetWorker provides support for the latest disk
backup and snapshot technologies.
NetWorker Hosts
The NetWorker
Server
Server supports
the backup and Node NetWorker
Node stores tracking Storage Node(s)
and configuration write data to
Client
information. and read data
Client
from devices.
NetWorker Client(s)
Client
generate the
backup data.
NetWorker Client
Server
Node Node
Client Client
y Performs recoveries
y Provides user interfaces for backup
and recovery
The NetWorker Client is the largest NetWorker software component and the fundamental host.
The client’s most important functions are to generate backups (also called save sets), push them
to a NetWorker storage node, and retrieve them during a recovery. NetWorker clients are
usually the data servers in an IT environment.
While performing a backup, the client also generates tracking information (including the name
of each file and directory backed up and the time of the backup) and sends it to the NetWorker
server where it is stored. This tracking information is used to facilitate point-in-time recoveries.
The client software includes graphical user interfaces (GUIs) and command-line utilities which
allow users to manually perform backup and recovery operations. NetWorker client software is
installed on all participating hosts in the data zone, including hosts that also perform the
functions of NetWorker server and NetWorker storage node.
Server
Node
Node
Client
Client
A NetWorker Storage Node is a host that physically controls a backup device and responds to
requests from the NetWorker server. The device may be either a direct-attached or SAN-
accessible device. If a device is controlled by a host other than the NetWorker server, it is
considered a remote device and the storage node controlling the device is referred to as a remote
storage node.
The NetWorker server is always a storage node and is the default storage node for backups.
Using remote storage nodes is optional, although they distribute the backup workload and can
reduce network traffic. Storage node hosts have both the NetWorker client and storage node
software installed.
During a backup, a NetWorker client sends backup data to a particular storage node based on the
client’s configuration. During a recovery, the client reads from the storage node that controls the
device containing the necessary volume.
Storage nodes also send tracking information, including details about save sets written to the
volume during a backup, to the NetWorker server. This information is used for future backups
as well as for recoveries.
NetWorker Server
CFI
CFI
Server CFI
Node Node
Media
Database
Resource Client
Client Directory
The NetWorker Server is the host that stores the configuration information, such as supported
clients and devices, when to run the backups, what data to back up, etc. The NetWorker server
also stores the online NetWorker databases that track the backups and volumes. These include
the Client File Indexes (CFI) and Media Database. The NetWorker server, as a client within the
data zone, automatically backs up the configuration information and tracking databases to
protect NetWorker data.
There is a single NetWorker server per data zone and it must be available for any NetWorker
activity to be supported in that data zone. NetWorker servers have NetWorker client, storage
node, and server software installed.
CFI
CFI
CFI
Server
/nsr/index/
Node
Media
Resource Database
Client Directory
/nsr/mm/
/nsr/res/
NetWorker control data is the collection of NetWorker configuration information and the
backup tracking information stored on the NetWorker server. NetWorker control data includes
three components:
y Resource Directory: This directory (the resource database) contains the resource files.
y Media Database: This database is used to track all volumes used by NetWorker and all
save sets written to those volumes.
y Client File Indexes (CFI): These databases track each file (pathname) backed up by
clients, allowing users to browse their backups for files from a
particular point in time. The NetWorker server creates and
maintains one CFI per physical client.
The next several pages provide more details on each NetWorker control data component.
00 01 02 03 04 05 06 07 08 09 00 01 02 03 04 05 06 07 08 09
Note: See Chapter 3, NetWorker Resources and Administrative Interfaces, and the
nsr_resource(5) man page for more information.
Important: NetWorker resource files are text files and are to be managed only through
NetWorker administrative interfaces. Do not edit them directly!!
Media
Database
The Media Database stores tracking information describing all save sets backed up within the
data zone and all volumes labeled by NetWorker. Save set information includes:
y Save set name
y Volume storing the save set
y NetWorker client that created the save set
y Date and time the save set was backed up
y Size of the save set
y Unique save set identifier (SSID)
y Aging status
y Backup level
Volume information includes:
y Volume name
y List of save sets on the volume
y Volume status
y Name of media pool that the volume belongs to
Note: Many more pieces of information are maintained in the media database. See the
mminfo(1) man page for more information.
CFICFI
CFI
A Client File Index (CFI) stores client-generated information about a client’s backed up files
and directories. For each file and directory that is backed up, a CFI contains its pathname, file
attributes such as permissions and ownership, and the time stamp of when the save set
containing the file was backed up.
NetWorker clients send tracking information to their CFI, which resides on the NetWorker
server. A single CFI exists for each physical client.
Node Node
NMC Server
Starting with NetWorker 7.3, administration of a NetWorker server is performed using the
NetWorker Management Console Server (NMC) GUI (commonly referred to as the Console
Server), a Java-based interface accessible from any supported web browser.
The Console server provides a global view of your NetWorker environment, allowing you to
centrally manage one or more NetWorker data zones. There will normally be only one Console
server in an enterprise, but multiple servers are allowed.
The Console server must have the NetWorker 7.3 (or later) client software already installed.
The Console server can generate a number of preconfigured reports using information gathered
from any or all of the NetWorker servers. Detailed customized reports can also be created.
Reporting is discussed in Chapter 12, Administering the NetWorker Management Console
Server.
nsrindexd
CFI
CFI
CFI
nsrd
Server
Resource nsrmmd
nsrmmd Directory nsrmmdbd
Node Media Node
nsrjobd DB
Client Client
nsrexecd nsrexecd
nsrexecd
Client
NetWorker daemon processes are involved in almost all NetWorker operations, including
backups and recoveries. There are one or more NetWorker daemons to support each of the three
NetWorker host functions:
y Client daemon
y Storage node daemon
y Server daemons
In a Microsoft Windows environment, the core NetWorker processes shown in the slide are
started via two NetWorker services.
The following pages provide more details about each of the core NetWorker daemons.
There are additional daemon processes associated with NetWorker Management Console.
These will be discussed shortly.
nsrexecd
Client
The NetWorker client daemon, nsrexecd (network save and recover execution daemon),
runs on NetWorker clients to support remote execution requests from NetWorker servers. For
example, nsrexecd executes a backup command at the request of the NetWorker server. The
nsrexecd process also determines which RPC ports to use to support and request NetWorker
services.
In a UNIX environment, nsrexecd is started automatically during system boot up. In a
Windows environment, nsrexecd is started via the NetWorker Remote Exec Service, which is
configured to start automatically during boot up.
The NetWorker Storage Node daemon, nsrmmd (network save and recover media
multiplexing daemon), runs on NetWorker storage nodes to support reading and writing of data
to devices. One nsrmmd is started for each device configured as a NetWorker resource.
The nsrmmd daemon:
y Receives backup data (save streams) from NetWorker clients.
y Writes the data to a volume in the backup device it is controlling.
y Can receive data from multiple backups concurrently, organizing the data into a special
NetWorker format which supports multiplexing.
y Sends tracking information to the NetWorker server to track data written to the volume.
y Reads data from the volume during operations such as recoveries and cloning.
nsrindexd CFI
CFI
CFI
Server nsrd
nsrmmd Resource
Node Directory nsrmmdbd Manages the Media
Database
nsrjobd Media
DB
Client
nsrexecd
Coordinates all scheduled backups, and manages
the resulting information and statistics.
The NetWorker server daemons provide access to NetWorker services such as configuration
information, support for backup and recovery requests, and access to the media database and
client file indexes. There are four NetWorker server daemons.
y nsrd - (network save and recover daemon) is the master daemon. nsrd manages the
NetWorker resource database, which contains almost all NetWorker configuration
information. It also starts the nsrmmdbd and nsrindexd processes. nsrd is
started automatically at system startup via a run-control script. Once started, nsrd
starts the other server daemons and one nsrmmd process for each device local to the
NetWorker server.
y nsrmmdbd - (network save and recover media management database daemon) provides
the read and write service for the media database.
y nsrindexd - (network save and recover index daemon) provides the read and write
service for the client file index databases.
y nsrjobd – (network save and recover job daemon) is responsible for coordinating all
scheduled backups. It stores information about these operations and provides it
to the NetWorker server and the NMC server for reporting purposes.
Note: See the nsrd(1), nsrmmdbd(1m), nsrindexd(1m), and nsrjobd(1m) man pages
for more information.
Reports
Data Zone Data Zone
lgto_gstdb
NW Server NW Server
gstd
Jobs NMC Server Jobs
DB gsttclsh dbsrv9 DB
Client
NetWorker Features
and
Backup Process Flow
y Ease of administration
y Efficient backups
NetWorker Advantages
Ease of administration Network-based backup protection
y GUI (Web-based browser using NMC) y WAN and SAN
y Command-line Interface (CLI) y Open-system protocols (TCP/IP, NDMP)
y Highly configurable y Cross-platform (UNIX, Windows, Linux, etc.)
Client
W2k3 SAN
UNIX
NetWorker
Data Zone Client Remote
Storage Node
© 2006 EMC Corporation. All rights reserved. EMC NetWorker Overview - 20
You can easily administer NetWorker from any host having a supported web browser.
NetWorker also provides many command-line utilities. NetWorker is highly configurable,
allowing you to customize backups to meet specific needs.
NetWorker provides network-based backup protection that functions in WAN and SAN
environments on open system protocols, including TCP/IP and Network Data Management
Protocol (NDMP).
NetWorker uses the client/server model, which distributes the workload and improves
performance. It is supported on the many operating systems listed in the EMC Legato Software
Compatibility Guide. The supported platforms seamlessly co-exist within the data zone. For
example, a Microsoft NetWorker client can back up to a UNIX NetWorker storage node and
vice versa.
NetWorker supports:
y Multiple concurrent backups
y Backups from different operating
systems
y Backups to multiple devices
Save Set 1 Storage
Save Set 2 Node
W2k3
UNIX
Save Set 3
Save Set 6
EMC NetWorker allows you to perform more efficient backups than most native operating
system backup software programs.
Many native solutions limit you to one backup at a time, one operating system, and only local
devices. NetWorker is not bound by these restrictions and is capable of backing up multiple save
sets from clients running different operating systems to any NetWorker-configured device.
Features that support these capabilities include:
y Multiplexing
y Open Tape Format
y Remote device support
These features are discussed on the following pages.
Data Rate
4 MB/s
Save Sets
4 MB/s
buffer
4 MB/s
4 MB/s
4 MB/s
4 MB/s Storage
4 MB/s
node Device/Volume
In an environment without multiplexing, only one stream of data is written to the device at any
given time. This situation is not ideal because as more clients perform simultaneous backups,
the tape drive’s throughput is not optimized.
24 MB/s
Save Sets
4 MB/s
buffer
4 MB/s
4 MB/s
4 MB/s
4 MB/s Storage
4 MB/s
node Device/Volume
Multiplexing enables the device to simultaneously write more than one save set to a volume.
This allows the device to write to the volume at the collective data rate of the save streams, up to
the maximum data rate of the device.
The amount of multiplexing allowed (the number of save sets that can back up simultaneously)
is primarily controlled by two NetWorker settings, server parallelism and device target sessions.
These settings will be discussed in detail in a later chapter.
UNIX
SS1
W2K
SS2
SS
OTF
1
SS2
Open Tape Format (OTF) is a data format that allows multiplexed, heterogeneous (UNIX,
Windows, NetWare, etc.) data to reside on the same tape.
As shown in the slide, clients wrap their backup data into packages containing platform-
independent data (save set chunks) before sending them to the storage node. The storage node
receives the packages and arranges them on a pallet (media records and media files). These
pallets are then stored in the warehouse, or tape volume.
The way the storage node organizes the pallets is also platform-independent (Open Tape
Format), allowing any NetWorker storage node to read the data. Because of Open Tape Format,
a NetWorker storage node can be migrated to a host running a different operating system.
Note: For more information on OTF, see the mm_data(5) man page or the white papers
available at software.emc.com.
The NetWorker
Server Server stores
tracking information. Storage
Node
CFI
Media
DB
Another feature of NetWorker that promotes efficient backups is the support of remote devices.
This means that data can be backed up to devices controlled by remote storage nodes. Recall
that a remote storage node is any storage node other than the NetWorker server. The
NetWorker server is the only local storage node. Devices controlled by the NetWorker server
are local devices. Devices controlled by remote storage nodes are remote devices.
This feature allows you to distribute the backup workload throughout the data zone.
Backup
Staging Recovery
(optional)
In addition to backup and recovery, NetWorker provides a full range of data protection functions
including tracking and reporting, staging, cloning and recovery.
Backup
A backup is a copy of data you want to protect by storing it in another location. NetWorker
provides a scalable solution to manage backups for a small network or an entire enterprise, and
enables you to automate and configure this process for speed and efficiency.
Recovery
Recovery is the process of restoring data to a given point-in-time. NetWorker makes this process
possible from anywhere within the NetWorker data zone. This feature allows users to retrieve
single files or restore an entire filesystem.
Aging
Aging determines the length of time that the backup data is available for recovery. NetWorker
allows you to specify:
y How long to keep the tracking information viewable (browsable)
y How long to protect each volume from recycling (overwriting)
Cloning
Cloning is the optional process of copying a save set to another storage volume.
Staging
Staging is the optional process of moving a save set from one storage volume (the physical
medium where the data is stored) to another.
4 1
nsrindexd
3 3
6 2
6
nsrmmd nsrjobd
1 5
nsrexecd
5 2
4
1
5
save
NetWorker Client
The slide shows a high-level process and data flow of a typical NetWorker scheduled backup.
1. The server’s nsrd starts a scheduled backup. nsrd asks nsrjobd to send a remote
execution request to the client’s nsrexecd, requesting that it execute the NetWorker save
command to perform the backup.
2. The save command started on the client communicates with the server’s nsrd (through
nsrjobd) to request backup support.
3. nsrd matches the backup to a storage node’s nsrmmd based on configuration information
and save request attributes.
4. Once the volume has been mounted on the backup device, nsrd directs the client to push its
data to the storage node.
5. The client:
- pushes the data to the storage node’s nsrmmd
- sends tracking information to its client file index (CFI) via the server’s nsrindexd
6. nsrmmd on the storage node:
- writes the data sent by the save command to the volume
- send tracking information to the media database via the server’s nsrmmdbd
7. When the backup is finished, the connections are dropped.
Chapter 1 - Summary
These are the key points covered in this chapter. Please take a moment to review them.