MSSQL - Student Resource Guide
MSSQL - Student Resource Guide
Copyright ©2015 EMC Corporation. All Rights Reserved. Published in the USA. EMC believes the information in this publication is
accurate as of its publication date. The information is subject to change without notice.
THE INFORMATION IN THIS PUBLICATION IS PROVIDED “AS IS.” EMC CORPORATION MAKES NO REPRESENTATIONS OR WARRANTIES
OF ANY KIND WITH RESPECT TO THE INFORMATION IN THIS PUBLICATION, AND SPECIFICALLY DISCLAIMS IMPLIED WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Use, copying, and distribution of any EMC software described in this publication requires an applicable software license. The trademarks,
logos, and service marks (collectively "Trademarks") appearing in this publication are the property of EMC Corporation and other
parties. Nothing contained in this publication should be construed as granting any license or right to use any Trademark without the
prior written permission of the party that owns the Trademark.
Acartus, Access Logix, AdvantEdge, AlphaStor, ApplicationXtender, ArchiveXtender, Atmos, Authentica, Authentic Problems, Automated
Resource Manager, AutoStart, AutoSwap, AVALONidm, Avamar, Captiva, Catalog Solution, C-Clip, Celerra, Celerra Replicator, Centera,
CenterStage, CentraStar, ClaimPack, ClaimsEditor, CLARiiON, ClientPak, Codebook Correlation Technology, Common Information Model,
Configuration Intelligence, Configuresoft, Connectrix, CopyCross, CopyPoint, Dantz, DatabaseXtender, Data Domain, Direct Matrix
Architecture, DiskXtender, DiskXtender 2000, Document Sciences, Documentum, elnput, E-Lab, EmailXaminer, EmailXtender , EMC2,
EMC, EMC Centera, EMC ControlCenter, EMC LifeLine, EMC OnCourse, EMC Proven, EMC Snap, EMC SourceOne, EMC Storage
Administrator, Enginuity, eRoom, Event Explorer, FarPoint, FirstPass, FLARE, FormWare, Geosynchrony, Global File Virtualization,
Graphic Visualization, Greenplum, HighRoad, HomeBase, InfoMover, Infoscape, Infra, InputAccel, InputAccel Express, Invista, Ionix,
ISIS, Max Retriever, MediaStor, MirrorView, Navisphere, NetWorker, nLayers, OnAlert, OpenScale, PixTools, Powerlink, PowerPath,
PowerSnap, QuickScan, Rainfinity, RepliCare, RepliStor, ResourcePak, Retrospect, RSA, the RSA logo, SafeLine, SAN Advisor, SAN Copy,
SAN Manager, Smarts, SnapImage, SnapSure, SnapView, SRDF, StorageScope, SupportMate, SymmAPI, SymmEnabler, Symmetrix,
Symmetrix DMX, Symmetrix VMAX, TimeFinder, UltraFlex, UltraPoint, UltraScale, Unisphere, VMAX, Vblock, Viewlets, Virtual Matrix,
Virtual Matrix Architecture, Virtual Provisioning, VisualSAN, VisualSRM, Voyence, VPLEX, VSAM-Assist, WebXtender, xPression, xPresso,
YottaYotta.
Why make all these changes? NMM simplification means improvements in backup and recovery
performance, removing complex and time-consuming maintenance of different architecture
layers, and removing the restrictions on workflows to enable new features such as block based
backups. We’ll take a look at each of these features in this course.
With NMM 9.0 simplification, NMM uses the VSS Common Requestor for all VSS framework
related operations and workflows, replacing Replication Manager. PowerSnap’s snapshot
management is no longer needed in NMM 9.0. NetWorker uses the traditional Save for SQL
Server and SharePoint VSS backups and the block based backup framework for Hyper-V and
Exchange server backups. VSS hardware providers are not supported.
You can see what a difference these changes have made in reducing complexity by comparing
the pre-NMM 9.0 diagram on the left with the NMM 9.0 diagram on the right.
With NetWorker 9.0, the media footprint is much less complex. There is a metadata record
that is linked to the save set records for the application data.
After creating a client resource, you create a protection group and add the client resource to
the group. Then, after creating a policy, you create a workflow in the policy and assign the
protection group to the workflow. The workflow specifies when and how often the actions in
the workflow will run. The number of backups taken per day is controlled by the schedule
times of the workflow.
Workflows enable you to run one or more actions either sequentially or in parallel. Actions
supported for backup operations are probe, check connectivity, backup and clone. A clone
action can be configured to occur after the backup action or it can be an action in a separate
workflow by using a save set or a query group type to specify the save sets to clone.
Here, we see an example of a backup workflow for a SQL VSS backup. In this example, the
client resource was added to a new group called SQLVSS which was associated with a new
workflow in the Bronze policy called SQLVSS. This workflow consists of one action, a backup
action. As part of the action configuration, a pool consisting of a Data Domain device was
chosen for backup storage.
SQL and SharePoint VSS backups are only supported for backup level Full. For those types of
backups, make sure to set the schedule in the backup action to all Full. Backups configured for
any other level will fail.
Details of the supported application versions and service packs are not shown here. For the
most up-to-date and accurate listing of hardware, operating system, service pack, and
application versions that the NMM client supports, please refer to the NetWorker Online
Software Compatibility Guide available on the EMC online support website.
Details of the supported application versions are not shown here. For the latest, detailed
information for supported application versions and OS configurations, please refer to the
NetWorker Online Software Compatibility Guide available on the EMC online support website.
For the latest information for system requirements and supported application and OS versions,
please refer to the NetWorker Online Software Compatibility Guide available on the EMC online
support website.
During SQL VDI transaction log backup operations, there are checks for existing full and
transaction log backups which are already saved for a database. This functionality is to ensure
there is a valid chain of SQL transaction logs, thereby enabling point-in-time recoveries. If a
log gap is detected, the backup is promoted to a full database backup. This is the default
behavior.
However, SQL log gap checking can take up to 5 to 30 seconds per database which can add up
to a significant amount of time in a backup of a SQL server with thousands of databases. With
NMM 9.0, the user has the option to turn off log gap detection to improve the performance for
SQL database backups.
To do this, enable the Turn off log gap detection during incremental backup option on
the client resource. By default, this option is not checked, that is, log gap detection is enabled.
This flag applies at both the database and SQL instance levels. NMM adds the flag to
application information for the backup
Note: If log gap detection is turned off, make sure there is no other backup product protecting
the SQL databases or SQL instances, including SQL native and DDBMA for SQL. When an
incremental backup is created by another backup product while log gap detection is turned off
in NetWorker, a log gap will be introduced; this will not be detected in NetWorker and the next
NetWorker backup will not be promoted to full. When a restore is performed, that backup
performed by the other product will not be included in the chain for restore because NetWorker
does not know about the backup.
• Copy only full backup of an AlwaysOn database on a secondary server, available only in the
federated workflow.
• Automatic determination of the preferred node of the AlwaysOn availability group and
spawning backup on that node.
As shown here, the Prefer Secondary option specifies that backups occur on a secondary
replica except when the primary replica is the only replica online. In that case, the backup
occurs on the primary replica. This is the default option.
Application information on the client resource for the cluster specifies the option
NSR_FEDERATED_BACKUP=yes.
Scheduled backups are performed using the traditional backup action within a data protection
policy workflow.
Note that using the “Join-only” option during recovery is not supported as there is no
incremental backup support with NMM SQL VSS.
If a SharePoint database is configured with Availability Group listener but not joined to the
AlwaysOn availability group, the database will be inaccessible from SharePoint after AlwaysOn
availability group failover happens to another node. In this case, after failover, the SharePoint
writer does not report this database.
First, NetWorker 9.0 server supports backups and restores of existing NMM 8.2.x and 3.0.x
clients.
Second, you can upgrade NMM 8.2.x and 3.0.x clients to NMM 9.0.
And third, NMM 9.0 can restore NMM 8.2.x and 3.0.x backups.
When the NetWorker server is migrated to version 9.0, the existing group resources are
migrated to NetWorker 9.0 protection policies, workflows, backup actions and groups. The
function of groups in 9.0 is now solely for selection of clients in workflows. Clients belong to
9.0 groups with the same group names as before migration. After migration, when a backup
containing a pre-9.0 client resource is initiated, the NetWorker server needs to determine
which values to use for pool, storage node, backup level and retention policy. This is done
based upon the value of the Client Override Behavior attribute of the backup action
associated with each client resource. Initially after migration, the Client Override Behavior is
set to Legacy Backup Rules. When using legacy backup rules, NetWorker uses the pool
specified in the client resource. If a pool is not specified in the client resource, it uses the pool
selection attributes in the pool resources. The values for storage node, backup level and
retention period are taken from the client resource. Where not specified in the client resource,
the value from the equivalent attribute in the backup action is used. This behavior remains in
effect until the Client Override Behavior attribute in the backup action is changed. Other
values for the attribute are Client Can Override, where values in the client resource for pool,
storage node, backup level and retention period take precedence over the values defined in
the backup action. If an attribute in the client resource is empty, then the value specified in
the equivalent attribute in the backup action is used. If the Client Override Behavior is set
to Client Can Not Override, then NetWorker uses the values for pool, storage node, backup
level and retention period specified in the backup action.
Because pre-9.0 NMM snapshot policies are not applicable in NMM 9.0, you may want to
manually rollover any existing snapshots before upgrading to NetWorker 9.0.
The NetWorker push upgrade feature can be used for standalone NetWorker clients for
NetWorker 8.1.1 or later and NMM 3.0 or later. Note that a direct upgrade cannot be
performed for NMM releases prior to NMM 3.0 and for clustered clients. To install without using
the push feature, first remove the existing client and NMM software and then perform an
install of NetWorker 9.0 client and NMM.
• For SharePoint, remove C:\inetpub and System Components from the Save set
attribute if they are present
• For VDI backups, clear the schedule information in the client resource and choose a
comparable new backup level and schedule in the backup option
Changing the contents of the Application Information field is not required as NMM 9.0 will
ignore any obsolete parameters in the Application Information field and log a message in
the application log noting the unsupported application information. An exception is the
NSR_VSS_FULL_Backup variable for Hyper-V client resources. In NMM 9.0, the behavior of
this variable has changed. To continue with the pre-9.0 behavior, you must specify
NSR_VSS_FULL_BACKUP=no in the Application Information field and backups will be
performed as Copy type backups. If the NSR_VSS_FULL_Backup variable is not specified in
the Application Information field or if the variable is set to “yes”, Hyper-V VM backups are
enabled (initial full, subsequent incrementals).
Note: You can use NMC bulk edit to change multiple client resources.
NMM 9.0 supports the restores of previous NMM 8.2.x and 3.0.x VSS backups through the
installation of an additional component available from the NMM 9.0 installer. Snapshot restores
and restores of Microsoft Data Protection Manager are not supported. A reboot of the NMM
host server is required after the installation.
To enable this functionality, the NetWorker 9.0 Extended client package, lgtoxtdclnt-
9.0.x.x.exe, must be installed first. Then, in the NMM 9.0 installer, select the Restore
previous NMM release backups option. This installs NMM 8.2.1 binaries and the Eldos driver
to support Granular Level Restores (GLR). On the wizard completing screen, select Restart to
complete the installation and reboot the host.
After the installation, there will be two shortcuts on the desktop for launching an NMM GUI:
one for restoring current backups and one for restoring previous backups. This component is
installed in a separate folder, rpnvmm, under \nsr; recover logs are generated in the
\nsr\rpvnmm\applogs folder.
Note that this component can be installed during the initial installation of NMM 9.0 or, if you
later realize that you need to install an older backup, you can install it then.
NMM 9.0 for Exchange software only backs up to AFTD or Data Domain devices and does not
support other device types.
The backups performed by NMM 9.0 for Exchange exclusively are VSS backups of the type
Exchange VSS full. VSS full backups truncate the transaction logs and other application log
files. This has the effect of making some third party incremental backups of the Exchange
server impossible causing all those backups to be full backups. The logs files are
truncated/deleted up to a checkpoint taken at the backup start time.
The first backup of an Exchange deployment gathers both the Exchange database files and the
Exchange application log files.
The Write Tracker driver maintains a bitmap, in memory, to track the changes to the blocks of
the database. If the system is rebooted, the Write Tracker driver must reinitialize and start a
new full backup of the database to create the block tracking bitmap in memory. This behavior
is by design.
CBT uses a map of all the blocks on a volume to track changes. The map is a set of bits where
every bit corresponds to a single block on disk. After a backup, all of the bits are set to 0. This
means that no changes have occurred yet. When a block on disk is changed, CBT updates the
map changing that block’s mapped bit from 0 to 1. CBT maps are retained between system
reboots.
NMM 9.0 for Exchange software will also force an incremental backup to a full backup under
several other conditions. These include:
When changing the target device where the backups are stored.
When the BBB framework promotion logic dictates that a promotion to a full backup is
required. This can happen when the system is rebooted, the WT detects that the WT bitmap
has been corrupted, the paths for the Exchange database or Exchange logs are changed, and
the volume size where the Exchange logs or the Exchange databases are located change size.
There are three tasks the Exchange Admin Configuration tool can perform: Configure Admin
User, Update Admin Password, and Validate an existing admin.
To use the Exchange Admin Configuration tool a user must be logged in with Domain
Administrator permissions.
Use the Configure Admin User function to create an NMM Exchange administrator account.
Supply the user name and password for the user and specify the database to be used by the
user. This creates an Active Directory NMM Exchange administrator account and configures the
appropriate permissions and memberships to perform Exchange database restore and granular
level recovery. This action will also create an Exchange security group called EMC NMM
Exchange Admin Roles.
Update Admin Password updates the password for an NMM Exchange administrator account
for backup and recovery where NMM is installed.
The Validate an existing user function verifies the configuration of the supplied user and
reports back if the user is correctly configured to run NMM Exchange administrator backups
and recoveries.
To make CBT possible for Hyper-V VMs, a new WMI property was introduced by Microsoft
called IncrementalBackupEnabled, which is enabled in the OS by default. Now, incremental
backups of VMs are possible. Using the incremental forever backup strategy, new VMs receive
a full backup first and subsequent incremental backups. Synthetic full backups are produced
from the combination of a full backup and incremental backups. Backup and recovery using
these methods result in quicker completion times.
NetWorker also takes advantage of Hyper-V 2012 and 2012R2 Partial Writer support. By
specifying this option in the client backup configuration, the backup of a large scale
deployment of VMs continues toward completion, even if an individual VM is found in a bad
state. The savegroup will show a status of passed despite any VMs that fail to create a
snapshot.
Data Domain and AFTD device types are the only supported backup devices.
Recoveries using NMM version 3.0.x or newer are only permitted if the Restore previous
NMM release backups option was selected during the NMM install.
Block based backups do not support data files that are encrypted, compressed or deduplicated.
For standalone block based backups, configure the Hyper-V client resources for Hyper-V server
backups. This can be done in the NetWorker Management Console. Specify the physical
hostname as the client. The additional protection elements, such as the policies and workflows,
must be configured specifically for Hyper-V. For example, an incremental forever strategy
should be chosen as the policy for Data Domain target devices. If using AFTD, it is best
practice to use “Incremental Forever, Full on 1st of Month.” Otherwise, a Full will be performed
automatically after the 38th incremental backup. This is to prevent long chains of differencing
disks in Hyper-V.
For clusters using CSV and federated backups, deployment is slightly different. The client
resource must be created for the cluster virtual server name instead of the physical
hostnames. The policy guidelines are the same as standalone servers. All VMs in the Hyper-V
cluster will be included in the save set. This includes any new VM clusters added to Hyper-V
later on. However, any standalone VMs in the same Hyper-V server as clustered VMs must be
managed as separate client resources.
In this simple architecture diagram, we have a three node cluster utilizing a virtual cluster
name as client identification. The virtual cluster name serves as a main proxy host and chosen
VMs perform as the virtual proxy hosts. The virtual proxies can work in parallel with physical
proxies to intelligently load balance all the CSVs during backups to ensure better backup
performance.
1. Start by creating a client for the virtual cluster name. This serves as the main proxy client.
Additional clients need to be created for each physical cluster host as well as the VMs that
will serve as the proxy hosts.
2. The virtual proxy hosts should be clustered VMs that are part of the same domain as the
physical hosts. NetWorker Module for Microsoft will need to be installed on each virtual
proxy host.
3. The main proxy client requires additional application attributes to be enabled in the client
configuration. Distribute the backup workload across the virtual proxy hosts by including
them in the Preferred Server Order List (PSOL). Also, ensure CSV ownership can change
during the backup operation to take advantage of intelligent load balancing. This ensures
CSVs are owned by the virtual proxy hosts for backing up VM files. By default, the
MOVE_CSV_OWNERSHIP option is set to Yes.
1. When a backup is started in a CSV environment, a list is generated of CSV ownership for
the group of VMs to be backed up.
2. Those CSVs are moved to the physical host, or coordinating node, where the different
proxies are running.
3. A separate save operation occurs to take snapshots of the CSV volumes. The snapshot can
occur from any cluster node.
4. The completed snapshot is mounted to the physical host where a corresponding CSV was
moved previously.
5. Each of the virtual proxy hosts start a data-rollover job from the shared snapshots for the
save sets assigned to them. This technique eliminates network traffic and maximizes
backup performance.
6. When the rollover is complete the cluster owner notifies the backup host.
The restore is a transparent operation. It has not changed from the previous version of
NetWorker.
• Follow the steps in the GUI, as reviewed on the previous slide, to locate the backup files
you wish to recover.
• When the files are found, ensure the Restore to a browser download location option is
enabled on the Restore Options page and click Finish.
• Once the restore process is complete, in the Restore Monitor, click the Download button
to save the recovered files. Choose the local folder or network path to save the recovered
files.
• The process of locating the files is the same as the browser download workflow.
• After the Restore Options page is reached, ensure the Restore to a browser download
location option is disabled.
• New choices are available in the GUI where the VM and destination file path can be
determined.
• Direct the restore to the VM and location for the recovered files and click Finish. Follow the
progress of the recovery using the Restore Monitor.
1. First, install the NetWorker client and then the NetWorker Module for Microsoft software on
the Hyper-V server.
2. Next, ensure the NetWorker Module for Microsoft Recovery Agent service is installed and
running. It should be located in the nsr\bin directory.
3. Finally, if you want to perform directed restores to a VM, perform the installation of the
Proxy Recovery Adapter on the VMs. The executable file will need to be copied from the
Hyper-V server to the VMs manually. The file is located at \EMC NetWorker\nsr\NMM
ProxyRA\x64 (or x86) on the Hyper-V host. Install the service from the command line
by running the executable with the “-add” option , then start the service from the
command line or Service Management . If you are only doing download restores, nothing
needs to be loaded on the VM.
5. In the Settings tab of FLR, specify a proxy URL . This will be the IP or FQDN of the server
where NetWorker Module for Microsoft is installed.
6. The list of Hyper-V servers will be available in the FLR GUI . Select the Hyper-V server to
display the VMs with available backups. You can now follow the steps as described in the
previous slides to recover files from the VMs.
This concludes the training. Please proceed to the course assessment on the next slide.