Dell PowerStore: Microsoft SQL Server Best

Practices
July 2022

H18250.7

White Paper

Abstract
This document includes architecture, deployment, and performance
guidance for Microsoft SQL Server and Microsoft Data Platform with
Dell PowerStore.

Dell Technologies
 Copyright

The information in this publication is provided as is. Dell Inc. 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 software described in this publication requires an applicable software license.
Copyright © 2020-2022 Dell Inc. or its subsidiaries. All Rights Reserved. Dell Technologies, Dell, EMC, Dell EMC and other
trademarks are trademarks of Dell Inc. or its subsidiaries. Intel, the Intel logo, the Intel Inside logo and Xeon are trademarks
of Intel Corporation in the U.S. and/or other countries. Other trademarks may be trademarks of their respective owners.
Published in the USA July 2022 H18250.7.
Dell Inc. believes the information in this document is accurate as of its publication date. The information is subject to change
without notice.

2 Dell PowerStore: Microsoft SQL Server Best Practices

 Contents

Contents
Executive summary ........................................................................................................................ 4

Introduction ..................................................................................................................................... 5

Scale-up and scale-out architecture.............................................................................................. 6

Host configuration .......................................................................................................................... 7

Sizing and performance optimization ........................................................................................... 8

Performance .................................................................................................................................. 18

Management and monitoring ....................................................................................................... 27

Data protection and disaster recovery ........................................................................................ 33

References ..................................................................................................................................... 36

Dell PowerStore: Microsoft SQL Server Best Practices 3 3

 Executive summary

Executive summary

Overview This paper provides best practices for using Dell PowerStore in a Microsoft SQL Server
environment. SQL Server is a robust product that can be used in various solutions. The
relative priorities of critical design goals such as performance, manageability, and
flexibility depend on the environment. This paper provides considerations and
recommendations to help meet design goals. For general best practices for using
PowerStore systems, see the document Dell PowerStore: Best Practices Guide.

These guidelines are intended to cover most use cases. They are recommended by Dell
Technologies, but they are not strictly required. For questions about the applicability of
these guidelines in your environment, contact your Dell Technologies representative to
discuss the appropriateness of the recommendations.

Audience This document is intended for IT administrators, storage architects, partners, and Dell
Technologies employees. This audience also includes any individuals who may evaluate,
acquire, manage, operate, or design a Dell networked storage environment using
PowerStore systems.

Revisions Date Description

April 2020 Initial release: PowerStoreOS 1.0

July 2020 Updated AppsON host configuration and sizing guidance

February 2021 Legal disclaimer update

April 2021 PowerStoreOS 2.0

May 2021 Clarification on PowerStore X and T

October 2021 Updated NVMe terminology

October 2021 Updated template

July 2022 PowerStoreOS 3.0

Note: This document may contain language from third-party content that is not under Dell
Technologies’ control and is not consistent with current guidelines for Dell Technologies’ own
content. When such third-party content is updated by the relevant third parties, this document will
be revised accordingly.

We value your Dell Technologies and the authors of this document welcome your feedback on this
feedback document. Contact the Dell Technologies team by email.

Author: Doug Bernhardt

Note: For links to other documentation for this topic, see the PowerStore Info Hub.

4 Dell PowerStore: Microsoft SQL Server Best Practices

 Introduction

Introduction

PowerStore and Dell PowerStore is a robust and flexible storage and compute option that is well suited to
SQL Server SQL Server. The capabilities of the PowerStore platform provide some unique benefits
and exciting architecture options for SQL Server workloads.

PowerStore PowerStore achieves new levels of operational simplicity and agility. It uses a container-
product based microservices architecture, advanced storage technologies, and integrated
overview machine learning to unlock the power of your data. PowerStore is a versatile platform with
a performance-centric design that delivers multidimensional scale, always-on data
reduction, and support for next-generation media.

PowerStore brings the simplicity of public cloud to on-premises infrastructure, streamlining

operations with an integrated machine-learning engine and seamless automation. It also
offers predictive analytics to easily monitor, analyze, and troubleshoot the environment.
PowerStore is highly adaptable, providing the flexibility to host specialized workloads
directly on the appliance and modernize infrastructure without disruption. It also offers
investment protection through flexible payment solutions and data-in-place upgrades.

Microsoft SQL Microsoft SQL Server is the industry-leading data platform. A product which once started
Server product as a database engine on Microsoft Windows has evolved into an entire data platform. This
overview platform enables cloud, hybrid, and on-premises deployments in several different
environments including Windows, Linux, containers, and Kubernetes. Besides the wide
number of environments that SQL Server can be deployed in, the use cases have
expanded. These uses now range from typical online transactional processing (OLTP) to
various analytics options such as data warehousing, artificial intelligence (AI), machine
learning (ML), and big data. The adaptable and flexible nature of SQL Server makes
PowerStore a perfect fit for SQL Server.

Prerequisite The best practices explained in this document require knowledge from the following
reading resources:
• PowerStore documentation on Dell.com/powerstoredocs
• Dell PowerStore Host Configuration Guide at Dell.com/powerstoredocs
• Microsoft SQL Server documentation library

Terminology The following terms are used with PowerStore.

Table 1. Terminology

Term Definition

Appliance Term used for solution containing a base enclosure and any
attached expansion enclosures. The size of an appliance could
be only the base enclosure or the base enclosure plus
expansion enclosures.

Base enclosure Used to reference the enclosure containing both nodes (node
A and node B) and the NVMe drive slots.

Dell PowerStore: Microsoft SQL Server Best Practices 5 5

 Scale-up and scale-out architecture

Term Definition

Cluster Multiple PowerStore appliances in a single grouping. Clusters

can consist of one appliance or more. Clusters are expandable
by adding more appliances. A cluster supports up to four
appliances.

Expansion enclosure Enclosures that can be attached to a base enclosure to

provide 25 additional SAS-based drive slots.

Node The component within the base enclosure that contains

processors and memory. Each appliance consists of two
nodes.

PowerStore Manager The web-based user interface (UI) for storage management.

Storage volumes (volumes) PowerStore volumes using block storage. These volumes are
displayed in the Block area of the PowerStore dashboard.

Virtual Volumes (vVols) PowerStore X model volumes support VMware vSphere Virtual
Volumes (vVols). These volumes are displayed in the
Virtualization area of PowerStore Manager.

Watchlist Configurable list of volumes or hosts that are available on the

dashboard for quick reference and monitoring.

Scale-up and scale-out architecture

Introduction Scale-up and scale-out are common architecture models used when accommodating a
growing data footprint. Microsoft SQL Server has supported both scale-up and scale-out
architecture models for several years. Dell PowerStore aligns with the SQL Server by also
providing both scale-up and scale-out architecture options. When designing data
architectures, there are tradeoffs in design, functionality, and scalability that must be
considered.

SQL Server When you scale up an instance of SQL Server, you can increase the amount of memory,
scale-up compute, and storage by upgrading to a higher software version or increasing the
hardware capabilities. You get the same or more functionality and retain the same logical
data design and therefore the data architecture is unaffected. This approach is common
because it is low impact.

SQL Server Scale-out architectures on SQL Server enable a data footprint to scale past what can be
scale-out achieved by a single instance. The trade-off is that these architectures require
considerably more planning and consideration and often require applications to be
designed accordingly. Concepts such as data placement, load balancing, and data
sharding may need to be addressed, and the impact on features such as replication or
availability groups.

PowerStore PowerStore models can scale up with several different upgrade paths. You can add drives
scale-up or drive expansion shelves to an appliance, add storage network paths, upgrade to next-
generation hardware as it becomes available, or upgrade to more powerful models within
the same family. Like SQL Server scale-up, this approach is also low impact. PowerStore

6 Dell PowerStore: Microsoft SQL Server Best Practices

 Host configuration

redundancy features and the Anytime Upgrades program offered by Dell Technologies
ensure a frictionless upgrade experience.

PowerStore PowerStore can also scale-out by adding nodes to the PowerStore cluster. PowerStore
scale-out can add up to four appliances in a cluster to provide extra capacity and performance. Like
scale-up or scale-out architecture decisions for SQL Server, there are similar
considerations for PowerStore. See the document Dell PowerStore: Clustering and High
Availability for complete details.

Host configuration

Introduction PowerStore provides the option of running workloads directly on the PowerStore
appliance as virtual machines or running workloads on external hosts and presenting
storage through block or file interfaces.

Before you create objects or deploy workloads on PowerStore, we recommend reviewing

the following guides for best practices and tuning recommendations. We recommend
applying these changes before deploying workloads since some changes require a reboot
of the PowerStore appliance nodes. For resiliency, ensure that external hosts are
connected through multiple paths, or multipath, to the PowerStore appliance.

External SQL PowerStore presents storage to external hosts through either block or file interfaces.
Server hosts Block storage is commonly used for SQL Server workloads due to the various speeds and
protocols that are offered which make it ideal for performance. With PowerStoreOS 2.0,
PowerStore supports 25 GbE and 32 Gb Fibre Channel (FC) and NVMe over Fibre
Channel (NVMe/FC). These options are recommended for SQL Server high-bandwidth
storage workloads such as analytics that can generate a large amount of large block I/O.
Instructions and best practices for configuring hosts can be found in the Dell PowerStore
Host Configuration Guide at Dell.com/powerstoredocs. File storage (SMB) can also be
used in environments running SQL Server on Windows.

For PowerStore T model appliances, follow recommendations for configuring external

hosts in the Dell PowerStore Host Configuration Guide. In addition, when running SQL
Server on virtual machines with VMware ESXi hosts, review the guidance in the
documents Dell PowerStore Virtualization Infrastructure Guide and Dell PowerStore:
VMware vSphere Best Practices.

Boot from SAN

PowerStore supports boot-from-SAN for environments that want to further virtualize
storage resources. In addition to the storage virtualization benefits, Boot-from-SAN can
also be used to protect the operating-system configuration and allow for quick recovery.
This configuration can be beneficial in environments where the recovery time objective
(RTO) is strict.

Dell PowerStore: Microsoft SQL Server Best Practices 7 7

 Sizing and performance optimization

AppsON Integration of the PowerStore container-based architecture with integrated VMware ESXi
results in a new level of consolidation for enterprise storage. This integration combines
the benefits of a local on-array application environment with unmatched integration with
the vSphere management environment and server resources. This ability allows users to
bring applications closer to storage, by running applications as virtual machines directly
on PowerStore.

Benefits of the AppsON capability include a new level of agility for application
deployments, with seamless movement between the PowerStore appliances and VMware
ESXi servers. AppsON also provides the ability to shrink the stack by eliminating server
and networking footprint for space-efficient edge and remote deployments. AppsON is
ideal for data-intensive applications that require low latency or a storage-heavy imbalance
of compute and storage.

AppsON can be an attractive option whether the goal is to optimize I/O performance by
reducing components in the data path, add performance to an existing vSphere cluster,
offload a noisy neighbor, or other reasons.

For PowerStore X model appliances, the PowerStore X Performance Best Practice

Tuning paper provides several tips that increase performance for SQL Server workloads.
These tips are especially helpful if you are running workloads that use block sizes greater
than 128k such as analytic workloads and backup/restore operations. Besides the
recommendations in this paper, when running workloads larger than 128k, you can find
other performance gains by increasing the iSCSI maximum I/O size, which is 128k by
default. To enable this result, run the following command on each host:

esxcli system settings advanced set -o /ISCSI/MaxIoSizeKB -i 512

Instructions for running esxcli commands on PowerStore X nodes are in the PowerStore X
Performance Best Practice Tuning paper.

Sizing and performance optimization

Introduction There are several best practices for provisioning compute and storage to SQL Server.
The size of power and flexibility of SQL Server allow it to be used in several different
deployment scenarios. This section covers some considerations for optimizing
PowerStore.

SQL Server The I/O storage system is a critical component of any SQL Server environment. Sizing
design and configuring a storage system without understanding the I/O requirements can have
considerations unfavorable results. Analyzing performance in an existing environment using a tool like
Live Optics can help define the I/O requirements. For best results, capture performance
statistics for a period of at least 24 hours that includes the system peak workload.

Demanding SQL Server workloads that require low latency, high bandwidth, or both can
benefit from moving the compute to PowerStore X model arrays. Moving these intensive
workloads onto the storage platform where compute and storage are located together can
deliver higher I/O performance for performance-sensitive applications. Also, virtual
machines running other application components can be on PowerStore X models for

8 Dell PowerStore: Microsoft SQL Server Best Practices

 Sizing and performance optimization

further performance gains. Using PowerStore X models in a larger vSphere cluster and
using VMware vSphere vMotion capabilities can be a powerful solution to balance SQL
Server workloads and application workloads across the entire environment.

Database design
When creating SQL Server databases, the default is to have one database file, one
filegroup, and one log file. If at any point the database will have high performance
requirements, it is recommended to use at least two data files per file group. SQL Server
will write data evenly across all files in the filegroup and therefore the underlying files.
Using multiple files allows the database workload to be spread across multiple volumes or
vVols. Even if they are placed on the same volume in the beginning, it is advantageous to
create this configuration from the start. More files can be added later, but the data must
be restriped to achieve balanced access through re-indexing or some other technique.
This reason is because SQL Server uses a proportional fill strategy when writing to data
files such that it will write more data to files that have a greater amount of free space.

More filegroups can be created and may be beneficial to use some SQL Server features.
However, the number of filegroups has no impact on performance from a storage
perspective.

SQL Server log files are accessed sequentially and do not use proportional fill. Therefore,
there is no performance benefit to creating extra log files.

OLTP workloads
While every environment is unique, an online transaction processing (OLTP) workload
typically consists of small, random reads and writes. A storage system that services this
type of workload is primarily sized based on capacity and the number of IOPS required.
PowerStore models allow the flexibility to be configured for block-optimized, unified (block
and file), or AppsON to meet unique OLTP requirements.

Analytic workloads
An online analytic processing (OLAP), decision support system (DSS), or big data
workload is typically dominated by large, sequential reads. A storage system that services
this type of workload is primarily sized based on throughput. When designing for
throughput, the performance of the entire path between the server and the drives in
PowerStore needs consideration. This is handled automatically when using virtual
volumes on PowerStore X models. When mapping volumes to hosts, consider using 32
Gb Fibre Channel (FC) or 25 Gbps iSCSI connectivity to the array. For high-throughput
requirements to be met, multiple HBAs may also be used in the server, the array, or both.

Mixed workloads
The most common scenario is a mixed workload environment. Typically, SQL Server I/O
patterns do not strictly fall into an OLTP or analytics pattern. This factor is what can make
SQL Server workloads challenging because no two workloads behave the same. In
addition, the same SQL Server host or instance may be servicing multiple applications or
transaction workloads. Mixed workload can also imply that multiple applications (in
addition to SQL Server) are residing on the same host or accessing the same storage.
The combined workload of these applications invalidates any typical application I/O usage

Dell PowerStore: Microsoft SQL Server Best Practices 9 9

 Sizing and performance optimization

pattern. For these reasons, it is important to gather performance metrics for best sizing
results.

SQL Server SQL Server is now available and supported on several different platforms. PowerStore
deployment presents several flexible options for various SQL Server deployments.
platforms
Windows
When deploying SQL Server on Windows on PowerStore, virtual volumes (vVols) and
storage volumes (volumes) are available on PowerStore X models. In addition to vVols
and volumes, PowerStore T also contains file support through SMB. All these options are
acceptable for use with SQL Server and rely on the configuration of the volumes through
Windows. When using SMB storage for SQL Server, ensure the Sync Writes Enabled and
Oplocks Enabled are turned on for the file system. See the Windows section of the
PowerStore Host Configuration Guide on the PowerStore Info Hub for instructions about
configuring host access on Windows.

SQL Server on Windows deployment

SQL Server on Windows is deployed with an executable (exe) file which can be found on
the Microsoft SQL Server website. Deployment is done by downloading the SQL Server
exe file for the intended version and running the executable. The SQL Server deployment
wizard will walk through several available options. Storage layout can either be
customized during the install or after install once the SQL Server instance is running.
When performing large enterprise deployments, there is an unattended install option
available to automate the installation.

Linux
When deploying SQL Server on Linux on PowerStore, vVols and volumes are available
on both PowerStore models. Although PowerStore T models contain file capabilities, they
do not support NFS v 4.2 which is the minimum that is required for SQL Server on Linux.
Therefore, NFS is not supported.

Be sure to follow Microsoft Installation guidance for deploying SQL Server on Linux for
supported Linux versions, minimum requirements, and general recommendations.
Currently, Red Hat Enterprise Linux, SUSE, and Ubuntu Linux distributions are supported.
The partnership of Dell Technologies, Microsoft, and RedHat provides a platform with full
end-to-end enterprise support. Therefore, Red Hat Enterprise Linux has quickly become a
platform of choice for deploying SQL Server on Linux.

SQL Server on Linux deployment

Once the Linux host is presented to the array, storage volumes can be presented to the
host. Since Linux is a relatively new platform for SQL Server, an example is provided
below for configuring a Red Hat Enterprise Linux host to consume PowerStore block
volumes. This example uses multipath, which is a best practice for resiliency. For
complete instructions about configuring Linux hosts and multipath, see the PowerStore
Host Configuration Guide on the PowerStore Info Hub.

After the storage volumes have been provisioned to the host, run the following command
to make the volumes visible to the host:

#/usr/bin/rescan-scsi-bus.sh –a

10 Dell PowerStore: Microsoft SQL Server Best Practices

 Sizing and performance optimization

Identify the volume name in the mapper folder by listing the multipath volumes and
searching for the volume WWN from PowerStore. The volume WWN is visible in the
PowerStore volumes list under Storage > Volumes. For example, if the WWN for the new
volume is 68ccf0980018e6c92364df1970e08d33, the following command will identify the
mapper folder to use:

# multipath -ll | grep 68ccf0980018e6c92364df1970e08d33

The volume will need to be formatted before use. See the Microsoft SQL Server
documentation for supported file systems. Currently, XFS and EXT4 are supported. In this
example, XFS is being used. For XFS, the mkfs.xfs command is used to format the
volume passing the correct mapper folder as the parameter.

# mkfs.xfs /dev/mapper/mpathdb

A mount folder needs to be created to represent the new volume. Create a folder to
represent the mount:

# mkdir /var/opt/mssql/data5

To make the mount persistent, the mapping needs to be stored in /etc/fstab so it will be
created at boot time. The entry in the file will consist of six fields contained on a single line
separated by a space: device, mount point, file system type, options, backup operation,
and file system check order. When mounting volumes for SQL Server, it is a Microsoft
best practice to add the noatime attribute to the list of options to disable updates to the
volume timestamp for better performance. The fields backup operation and file system
check order will be left at the default of 0 and can be changed as needed. Based on our
example, the new entry in the /etc/fstab file would look like this:

/dev/mapper/mpathdb /var/opt/mssql/data5 xfs defaults,noatime 0 0

Once the entry for the new mount is created, running the mount command will read the
file and refresh the mounted devices. Even if the mount was manually created, it is a good
idea to run the mount command after updating /etc/fstab to ensure there are no errors. If
there are errors in the file, it could prevent Linux from booting properly.

# mount -a

Finally, before SQL Server can access the new mount, permissions need to be set. In a
production environment, follow the security best practices for your organization when
assigning permissions and ownerships to mount points. This example uses the same
default permissions and ownership from the SQL Server installation by copying the
permissions from the /var/opt/mssql/data folder and applying them to the new mount
point:

# chmod --reference /var/opt/mssql/data /var/opt/mssql/data5

At this point, the new volume is ready for SQL Server to use. See the Linux section of the
PowerStore Host Configuration Guide on the PowerStore Info Hub for further instructions
on configuring storage volume access on Linux.

Dell PowerStore: Microsoft SQL Server Best Practices 11 1

 Sizing and performance optimization

Containers
Running SQL Server inside Docker containers is a deployment option that was introduced
with SQL Server 2017. By default, the database storage is inside the container and is
ephemeral in nature. If databases need to be persisted beyond the lifetime of the
container, external storage must be presented. External storage is presented to a
container through a bind mount which is a local file system folder or mount from the
operating system that is presented to the container at startup. Follow the storage
guidance above for either Windows or Linux to create file system folders or mount points
to present as a volume bind mounts to containers.

SQL Server container deployment

Deploying SQL Server inside Docker containers is simple. Below is an example of running
SQL Server inside a container followed by a brief explanation of the parameters:
# sudo docker run -e "ACCEPT_EULA=Y" -e "SA_PASSWORD=P@ssword1" -p
1401:1433 -v /root/sql1/tpcc_data:/var/opt/mssql/data2 -v
/root/sql1/tpcc_log:/var/opt/mssql/log2 --name sql1 -d
mcr.microsoft.com/mssql/server:2019-latest

Table 2. Docker run parameters

Parameter Description

-e Sets environment variables, in this case options for licensing and the SA
password are set

-p Publishes the SQL Server port 1433 inside the container as 1401 on the host

-v Bind mount for a volume. In this example, three are specified. The format is
local mount:container mount
--name Name of the container, otherwise it will be assigned a random identifier

-d Container image to run. If it does not exist, it will try to locate it from any
configured repositories and download it

Kubernetes
Due to the clustered architecture of Kubernetes, there are special storage considerations
when running SQL Server instances or SQL Server Big Data Clusters on Kubernetes.
Kubernetes has created its own method of dealing with persistent storage in a Kubernetes
cluster using persistent volumes.

Dell Container Storage Modules (CSM) for PowerStore enable integration and automation
of PowerStore storage resources to Kubernetes. PowerStore storage integrates with
Kubernetes through the PowerStore CSI driver. The PowerStore CSI driver supports
storage volumes only. Kubernetes clusters running on PowerStore X models can either
use the vSphere Cloud Provider for virtual volumes or the PowerStore CSI driver for
storage volumes. The PowerStore CSI driver and CSI drivers for other Dell Technologies
products can be found on GitHub.

Whenever possible, perform storage operations through Kubernetes. In some cases,

modifying CSI volumes directly on the PowerStore appliance can cause CSI operations to

12 Dell PowerStore: Microsoft SQL Server Best Practices

 Sizing and performance optimization

fail or volumes to be orphaned. For example, CSI volumes added to a volume group need
to be removed from the volume group before the CSI driver can delete them.

In addition to the CSI driver support for PowerStore, there are PowerStore CSM modules
available for observability and replication. The PowerStore CSM Observability module
provides storage performance metrics to Kubernetes that are stored in Prometheus and
can be viewed with tools such as Grafana. This allows full end-to-end visibility of storage
metrics without the involvement of the storage administrator.

Figure 1. PowerStore volume topology in Grafana

Figure 2. PowerStore volume performance metrics in Grafana

Dell PowerStore: Microsoft SQL Server Best Practices 13 1

 Sizing and performance optimization

Figure 3. PowerStore read IOPS by volume in Grafana

The PowerStore CSM Replication module allows control of array-based replication

features from the K8s control plane. Volume can be replicated and failover and reprotect
operations can be performed.

The capabilities of the Kubernetes CSI specification are constantly evolving to support
new enterprise storage features. Use the latest Dell Technologies PowerStore CSI driver
and keep up to date on new releases. The latest documentation for Dell Technologies
CSM modules for all Dell storage products can be found at https://dell.github.io/csm-
docs/docs/.

SQL Server Kubernetes deployment

PowerStore storage for SQL Server running on Kubernetes is deployed and managed
almost entirely through Kubernetes. The PowerStore CSI driver translates Kubernetes
storage provisioning commands into PowerStore storage provisioning commands that are
orchestrated on the storage appliance. Since common storage provisioning tasks such as
creating, deleting, and expanding volumes are done through Kubernetes, the DBA or
Kubernetes administrator can provision storage without knowing the details of PowerStore
or requiring direct access. This ability enables self-service of common storage
provisioning tasks.

Once the PowerStore CSI driver is deployed, one or more Kubernetes storage classes will
be created. These storage classes describe the details of the CSI driver, encapsulate the
connectivity and protocol options of the storage, and contain defaults for common storage
provisioning values. Below is an example of powerstore-xfs storage class definition:

allowVolumeExpansion: true
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
annotations:
meta.helm.sh/release-name: powerstore
meta.helm.sh/release-namespace: csi-powerstore
creationTimestamp: "2020-10-07T13:33:14Z"

14 Dell PowerStore: Microsoft SQL Server Best Practices

 Sizing and performance optimization

labels:
app.kubernetes.io/managed-by: Helm
name: powerstore-xfs
resourceVersion: "221004"
selfLink: /apis/storage.k8s.io/v1/storageclasses/powerstore-xfs
uid: cd013881-522e-4e16-bbd0-1f3c31315c94
parameters:
FsType: xfs
provisioner: csi-powerstore.dellemc.com
reclaimPolicy: Delete
volumeBindingMode: Immediate

Kubernetes storage is provisioned as Persistent Volumes and Persistent Volume Claims.

More information about Persistent Volumes and Persistent Volume Claims can be found
at Kubernetes.io. For example, to create three volumes for a SQL Server pod running on
Kubernetes, the definitions are stored in a YAML file:

# cat pvc.yaml
kind: PersistentVolumeClaim
apiVersion: v1
metadata:
name: mssql-data
annotations:
volume.beta.kubernetes.io/storage-class: powerstore-xfs
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 8Gi
---
kind: PersistentVolumeClaim
apiVersion: v1
metadata:
name: mssql-data2
annotations:
volume.beta.kubernetes.io/storage-class: powerstore-xfs
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 5Gi
---
kind: PersistentVolumeClaim
apiVersion: v1
metadata:
name: mssql-log2
annotations:
volume.beta.kubernetes.io/storage-class: powerstore-xfs
1

Dell PowerStore: Microsoft SQL Server Best Practices 15 1

 Sizing and performance optimization

spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 2Gi

To create the persistent volume claims, the YAML file above is used as input. Running the
following command will create the volumes on PowerStore and create a volume alias that
can be used when creating the SQL Server pod.

#kubectl create -f pvc.yaml

Next, to deploy SQL Server a definition file is also used. The persistentVolumeClaim
attributes in the definition file refer to the Persistent Volume Claims created in the
previous step.

#cat sql.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
name: mssql-deployment
spec:
replicas: 1
selector:
matchLabels:
app: mssql
template:
metadata:
labels:
app: mssql
spec:
terminationGracePeriodSeconds: 30
hostname: mssqlinst
securityContext:
fsGroup: 10001
containers:
- name: mssql
image: mcr.microsoft.com/mssql/server:2019-latest
ports:
- containerPort: 1433
env:
- name: MSSQL_PID
value: "Developer"
- name: ACCEPT_EULA
value: "Y"
- name: SA_PASSWORD
valueFrom:
secretKeyRef:
name: mssql
key: SA_PASSWORD

16 Dell PowerStore: Microsoft SQL Server Best Practices

 Sizing and performance optimization

volumeMounts:
- name: mssqldb
mountPath: /var/opt/mssql
- name: mssqldata
mountPath: /var/opt/mssql/data2
- name: mssqllog
mountPath: /var/opt/mssql/log2
volumes:
- name: mssqldb
persistentVolumeClaim:
claimName: mssql-data
- name: mssqldata
persistentVolumeClaim:
claimName: mssql-data2
- name: mssqllog
persistentVolumeClaim:
claimName: mssql-log2
---
apiVersion: v1
kind: Service
metadata:
name: mssql-deployment
spec:
selector:
app: mssql
ports:
- protocol: TCP
port: 1433
targetPort: 1433
type: NodePort

To deploy SQL Server inside of a Kubernetes pod, first a secret is created to store the
password:
#kubectl create secret generic mssql --from-
literal=SA_PASSWORD="P@ssword1"

Next, using the YAML file above as input, kubectl create will deploy the SQL Server
instance on Kubernetes:
# kubectl create -f sql.yaml

Azure Arc-enabled SQL Managed Instance

Arc-enabled SQL Managed Instance (MI) allows SQL Server to be deployed from the
Azure portal to an on-premises Kubernetes cluster. Since Arc-enabled SQL MI deploys an
availability group, shared (read-write many) storage is required in addition to block
volumes. PowerStore with combined block and file is uniquely suited to run these
workloads on a single appliance. Dell Technologies and Microsoft performed a joint
performance and scalability study on running Arc-enabled SQL MI on PowerStore.
Complete details can be found in the paper Dell PowerStore with Azure Arc-Enabled Data
Services.
1

Dell PowerStore: Microsoft SQL Server Best Practices 17 1

 Performance

RAID Since the inception of storage devices that used multiple drives for a single volume, the
configurations topic of RAID configuration and storage pools has been highly debated. Designing and
and storage maintaining the proper design could be time consuming and the wrong decisions could
pools have a negative impact on performance. In addition, reconfiguring the storage layout often
required considerable planning.

PowerStoreOS uses the capabilities of flash-based storage for maximum performance

and capacity, eliminating the need for administrators to configure RAID or storage pools.
Most existing SQL Server publications that discuss RAID levels and configuration were
written based on older technology and apply to spinning drives. Therefore, these concepts
do not apply to PowerStore. Dell Technologies recommends following the guidance in this
paper and associated publications for best performance and availability. For further
questions, contact a Dell Technologies representative.

Validating the I/O Before deploying workloads on PowerStore, it is a good idea to validate the I/O path
path between the server and the array. Running a large block sequential read test using small
files should saturate the path between the server and the array. This test verifies that all
paths are fully functional and can be used for I/O traffic. Run this test on an isolated
server and PowerStore. Running this test in a production environment could cause
significant performance issues for active workloads. To validate the I/O path, run a large
block sequential read test using the following guidelines:
• Create one volume per node.
• Format the volumes using a 64 KB allocation unit.
• Use a block size of 512 KB for the test.
• Configure the test for 32 outstanding I/Os.
• Use multiple threads; eight is the recommended starting point.
If the displayed throughput matches the expected throughput for the number of HBA ports
in the server, the paths between the server and PowerStore are set up correctly.

MPIO A best practice for SQL Server hosts is to provide multiple paths to the storage for
resiliency and performance. This practice protects against possible data loss or downtime,
should a physical component fail. In scenarios where multiple paths to the storage are
used, MPIO must be enabled and configured. See the document Dell PowerStore Host
Configuration Guide at Dell.com/powerstoredocs for recommendations.

Performance

Overview Managing storage performance with PowerStore is simple and intuitive. The performance
policy can be adjusted at the volume or vVol level by selecting high, medium, or low.
Performance policies are relative in PowerStore, so leaving all volumes at the default of
medium provides equal performance to all volumes. The volumes all perform at the
highest level until a PowerStore volume is set to a different performance policy.

Selecting a performance policy sets a relative priority for the volume using a share-based
system such that when resources are constrained, the resource with the higher

18 Dell PowerStore: Microsoft SQL Server Best Practices

 Performance

performance policy receives priority. If there is no PowerStore resource constraint, the

volumes perform equally.

PowerStore In some cases, SQL Server can be very storage-resource intensive. For I/O intensive
Node Affinity SQL Server databases, multiple volumes are recommended. This use allows the flexibility
to balance the workload across both nodes in a PowerStore appliance. The “ownership” of
a volume by a node is known as Node Affinity.

PowerStore performs best when the system performance is balanced across both nodes
in the appliance. The Node Affinity is set when the volumes are mapped to a host. By
default, volume Node Affinity alternates between the two nodes in the appliance when
mapping volumes to a host to balance the workload. Usually, this process sufficiently
balances the load across both nodes in the appliances. Since every SQL Server workload
is unique, and storage access patterns depend on database layout across one or more
files, Node Affinity adjustments may be necessary to maximize overall system
performance.

Overall system and individual appliance performance can be viewed in the Hardware
section on the Performance tab. Scrolling down in the view will reveal performance by
node.

Figure 4. PowerStore hardware performance

Ideally, each node in the appliance will have equal performance. If the workload is
unbalanced between the nodes, you can evaluate the Node Affinity of the workload and
adjust.

Note: Node Affinity can only be adjusted on block volumes, vVols are system select only.

Dell PowerStore: Microsoft SQL Server Best Practices 19 1

 Performance

Volume node affinity is shown in the Volumes list under Storage. By default, this column is
not displayed. To add it to the view, click the Columns selection icon next to the filter icon.
Scroll down the list, and select Node Affinity.

Figure 5. Selecting additional columns in the Volumes list

Figure 6. Volumes list with Node Affinity added

“System select node A” and “System select Node B” are the system assigned values.
Node Affinity can be set using the PowerStore CLI (CLI) or PowerStore REST API. This
example uses the CLI.

The following command will display the details of a volume named “SQLData2-007’:

pstcli -destination <PowerStore Manager URL/IP> -u <id> -p

<password> /volume -name "SQLData2-007" show

To make changes, the volume id is required. You can either find the volume ID from the
above command or append “ -select id” to return only the volume id:

pstcli -destination <PowerStore Manager URL/IP> -u <id> -p

<password> /volume -name “SQLData2-007” show -select id

Note: The “id” used here is not displayed in the PowerStore Manager and therefore needs to be
discovered.

After the volume ID is discovered, it can be used in the following command to set the
Node Affinity. The following command sets the Node Affinity for volume a5edff43-e9a7-

20 Dell PowerStore: Microsoft SQL Server Best Practices

 Performance

48bd-9b19-e46c7d87d7d2 to Preferred_Node_A. Valid values are Preferred_Node_A,

Preferred_Node_B, and System_Select_At_Attach.

pstcli - destination <PowerStore Manager URL/IP> -u <id> -p

<password> /volume a5edff43-e9a7-48bd-9b19-e46c7d87d7d2 set -
node_affinity “Preferred_Node_A”

Note: When Node Affinity is set to Preferred_Node_A or Preferred_Node_B, it is no longer

managed by PowerStore unless it is set to System_Select_At_Attach.

Virtualization When running SQL Server inside virtual machines (VMs), it is recommended to start with
the VMware document SQL Server on VMware Best Practices Guide for performance and
scalability. This guide provides guidance for configuring SQL Server, the operating
system, and virtual machines for best performance on VMware.

For more tuning recommendations and best practices regarding VMware vSphere,
VMware ESX, and PowerStore T or PowerStore X (AppsOn), see the documents Dell
PowerStore: Virtualization Integration and Dell PowerStore Virtualization Infrastructure
Guide at Dell.com/powerstoredocs.

In addition to these general best practices for virtualization, there are some VMware
features and integrations for SQL Server.

VMware vVols vs VMDKs

Snapshots are a great way to protect SQL Server virtual machines and are encouraged
whenever possible to provide an additional recovery point to SQL Server backups.
However, taking snapshots on VMDK disks can result in a decrease in performance and
sometimes lengthy maintenance procedures to maintain snapshot copies, coalesce, and
delete.

vVol snapshots are offloaded to PowerStore, using hardware snapshot capabilities. This
process removes the traditional issues of VMDK snapshot performance and maintenance.
There is no performance penalty for vVols that contain snapshots. In addition, there are
no lengthy maintenance procedures.

Also, Dell Technologies has tested the performance of VMware vVols and VMDKs. In our
SQL Server workload testing, the performance has been proven to be equal on volumes
without snapshots. On volumes that contain snapshots, the performance on vVols is
superior. Therefore, vVols are recommended when running SQL Server workloads to
provide superior snapshot performance.

A complete discussion of vVols vs VMDK disks is in the VMware article Understanding

Virtual Volumes.

Metro Volume
PowerStore Metro Volume allows synchronous replication of VMFS datastores in a
vSphere Metro Storage Cluster (vMSC). This enables vSphere stretched cluster
capabilities within or across sites. It can also be used to accelerate VM migrations. One of
the challenges with migrating SQL Server virtual machines can be the time it takes to

Dell PowerStore: Microsoft SQL Server Best Practices 21 2

 Performance

perform a storage migration due to data size. Metro Volume can be used to accelerate
migrations between sites, in addition to high availability and fault tolerance. For complete
details on PowerStore metro volume, see the document Dell PowerStore: Metro Volume.

AppsON When running workloads on virtual machines on the PowerStore X appliance or AppsON
mode, there are some important considerations when sizing workloads.

General VMware tuning and resource allocation principles apply; consult the VMware
article About vSphere Resource Management for general tuning guidelines. If resources
in the vSphere cluster become a bottleneck, guest storage performance will suffer,
reducing the effectiveness of running workloads in AppsON mode.

Resource reservation
PowerStore X appliances reserve approximately half the CPU and memory resources for
running PowerStore appliance tasks. When sizing workloads to run on the appliance, this
needs to be considered. This reservation is visible in vSphere as the CPU and memory
will be displayed as consumed by the PowerStore X controller virtual machines. Alsothe
ESXi cluster cannot run at 100% capacity without incurring degraded performance. A
minimal amount of resources are required for ESXi overhead. See the VMware articles
Configuring Resource Allocation Settings and Administering Memory Resources for
resource reservation and sizing guidance.

For production applications where guaranteed performance is necessary, make sure to

size for possible node failover in the PowerStore cluster.

SQL Server virtual machines

General best practices apply when running SQL Server virtual machines in AppsOn
mode, but there are two items to note. First, VMware vVols will typically be used as
storage for VMDKs. While other storage options are still available, PowerStore X is
optimized for vVols storage, so this option should be the first choice. Also, use the
VMware Paravirtualized SCSI (PVSCSI) Controller as the virtual SCSI Controller. The
PVSCSI controller improves performance and reduces CPU overhead.

For more best practices guidance for running SQL Server virtual machines, see the
VMware guide Architecting Microsoft SQL Server on VMware vSphere.

Multi-appliance virtual machine migrations

Starting with PowerStoreOS 2.0, up to four PowerStore X appliances can be added to a
PowerStore X cluster. Then, AppsOn workloads can be moved among appliances in the
PowerStore X cluster using VMware vSphere vMotion. When moving virtual machines
among appliances in an AppsOn cluster, the storage will not migrate since PowerStore
Storage Containers are presented as a vVols datastore across all appliances in the
cluster. However, the vVols do reside on a single appliance. Therefore, this must be
considered when migrating VMs between clustered appliances. For a VM, running the
compute on one appliance when the vVols reside on another appliance will result in
degraded performance.

22 Dell PowerStore: Microsoft SQL Server Best Practices

 Performance

In the Compute-> Virtual Machines view below is a VM running on PS-9-appliance-1-

node-B.

Figure 7. Virtual Machine residence

Examining the vVols that belong to this VM shows that they are on PS-9-appliance-2.
These vVols should be migrated to PS-9-appliance-1. This process is done by selecting a
volume and then selecting Migrate. This action can also be done in the PSTCLI using the
migration_session create command.

Figure 8. vVol residence

Once the migration has been started, it can be viewed in the Migration section.

Figure 9. vVol migration

Dell PowerStore: Microsoft SQL Server Best Practices 23 2

 Performance

SQL Server Big Data Clusters

The Kubernetes platform performs resource balancing and scheduling within the Big Data
Cluster and is not aware of the virtualization layer. When optimizing for performance, it is
recommended to balance the workload across all nodes in the AppsON environment. In a
simple scenario, there would be an even number of worker nodes in the Big Data Cluster.
More complex scenarios may require placing certain nodes in the AppsON environment
and using Kubernetes scheduling features to place targeted workloads on AppsON
nodes. These concepts are covered in the Kubernetes documentation Assign Pods to
Nodes.

Kubernetes nodes in the master role consume relatively few resources. However, they
need to be responsive and highly available. Master node VMs could be run on other
available nodes in the vSphere cluster when applicable to reserve resources for the
AppsON environment.

Regardless of where master role VMs run, based on internal testing, it is recommended to
reserve CPU and memory resources for these K8s masters. If resource contention is
detected within K8s masters and they cannot immediately get the resources they require,
it can panic the K8s resource monitor and workload pods may be restarted.

Reducing SQL Optimizing SQL Server workloads will improve the performance and scale of the storage
Server I/O layer and ultimately the related applications. The following sections include common
techniques to reduce SQL Server storage I/O.

Memory
Allocating the proper amount of memory to SQL Server can increase performance and
avoid unnecessary I/O. SQL Server performs all I/O through the buffer pool (cache) and
uses a large portion of its memory allocation for the buffer pool. Ideally, when SQL Server
performs I/O, the data is already in the buffer pool and it does not need to go to disk. This
type of I/O is referred to as logical I/O and is the most desirable because it results in the
best performance. If the SQL Server data does not need to reside in the buffer pool, it
needs to access disk resulting in physical I/O. Proper memory allocation is critical to SQL
Server performance and can improve storage performance as well. Often, SQL Server
and storage performance can be further improved by adding additional memory to the
SQL Server instance. Generally, allocating more memory is better, but there is a point of
diminishing returns that is unique to each environment.

Buffer pool extension

Starting with SQL Server 2014, the buffer pool can be extended to a file on the file system
to provide additional space to cache data or index pages. Using this feature can provide
significant performance benefits without adding memory to the database server in some
cases. With more pages cached on the server, the I/O load on the array is reduced. When
placing the buffer pool extension on the array, create a separate volume for the buffer
pool extension and do not configure snapshots for the buffer pool extension volume. The
buffer pool data is repopulated by SQL Server when the instance is restarted. Therefore,
data recovery does not apply.

24 Dell PowerStore: Microsoft SQL Server Best Practices

 Performance

Persistent memory
SQL Server 2016 introduced support for persistent memory (PMEM). The capabilities of
PMEM are being expanded with SQL Server 2019 to cover more scenarios and support
the Linux operating system. Often, PMEM can be used to accelerate challenging I/O
workloads and make I/O patterns more efficient with SQL Server PMEM features such as
tail-of-log-cache and Hybrid Buffer Pool. Virtualized environments running Hyper-V or
VMware can also take advantage of PMEM making it a wise investment. Dell servers
support PMEM starting with Dell PowerEdge 14th-generation servers. For more
information, see Storage Class Memory in SQL Server 2016 SP1 and the Dell NVDIMM-N
Persistent Memory User Guide.

Instant File Initialization

By default, SQL Server writes zeros to the data file during the allocation process. The
process of zeroing out the data files consumes I/O and acquires locks as the SQL Server
data pages are written. This activity can occur for minutes or even hours depending on
the file size. While this result may seem minor, writing zeros to these files can occur at
disruptive periods when time and performance are critical. Example scenarios include
database auto growth, expanding a full data file, replication, or restoring a database as
part of a disaster-recovery event. When Instant File Initialization is enabled, SQL Server
will skip the process of zeroing out its data files when allocating space. Dell Technologies
recommends enabling Instant File Initialization.

Resource Governor
The Resource Governor was added in SQL Server 2008 to allow database administrators
to limit the CPU and memory resources a SQL query can consume. This feature was
enhanced in SQL Server 2014 to allow I/O resources to be limited as well. For example,
the Resource Governor can be used to reduce the impact of a user running an I/O
intensive report by limiting the maximum number of IOPS that user can perform. While a
query throttled by the Resource Governor takes more time to complete, overall database
performance is better.

Database design considerations

Reducing SQL Server I/O requires a holistic approach. Many of the items in this section
require involvement from the entire team responsible for SQL Server applications. This
team could include the business owner, architect, developer, database administrator, and
system administrator. Decisions at the design level have a multiplied downstream impact
as data is written and read multiple times and duplicated in various types of database
copies. These include databases that are copied for other uses such as testing and
reporting, replicated databases, replicated storage, and backups. One of the most
challenging aspects of SQL Server is that the I/O pattern and the amount of I/O that is
generated can vary greatly depending on the application, even if those applications have
databases of the same size. This is because the design of both the database and the data
access code control SQL Server I/O.

Database tuning can be one of the most cost-effective ways to reduce I/O and improve
scalability. At a high level, consider the following areas when tuning a database to reduce
I/O.

Dell PowerStore: Microsoft SQL Server Best Practices 25 2

 Performance

Database design
The foundation of the entire database and the schema for how data will be stored and
ultimately accessed is determined by the database design. The database design should
support both usability and efficient data access. This includes efficient table design and
data types as well as indexes, partitioning, and other features that can improve efficiency.
It is common for database design to only be focused on usability while performance and
scale are overlooked.

Query design
How a query is written can greatly affect the amount of I/O SQL Server must perform
when running the query. Queries should return only the required amount of data in the
most efficient manner possible. Tune the queries responsible for consuming a relatively
large number of resources as well as possible for best performance and scale.

Application design
Consider how applications are using the data and the way it is requested. Sometimes
code and component reuse can result in the same data being unnecessarily retrieved
repeatedly. All data access should be purposeful.

Maintenance
SQL Server uses a cost-based optimizer to generate query plans for data access. These
plans are based on the statistics regarding how data is distributed in the tables. If the
statistics are inaccurate, bad query plans may result and unnecessary I/O will be
performed. Proper database maintenance includes ensuring that statistics are up to date.
Frequent data modifications can also lead to fragmentation within SQL Server data files,
producing unnecessary I/O. Fragmentation can be addressed through index
reorganization or rebuilds as part of regular database maintenance. The database
maintenance process itself can also have a large I/O impact. Typically, every table and
index does not need to be rebuilt or reorganized every time maintenance is run. In
addition, table partitioning strategies can also be leveraged to make the maintenance
process more selective. Consider implementing intelligent maintenance scripts that utilize
usage statistics to perform maintenance on an as-needed basis. For mission-critical
databases, maintenance activities need to be considered as part of the overall design. If
maintenance is not considered as part of the overall process, issues can arise, such as
unmanageable sizes and feature incompatibilities that limit available options and
strategies.

26 Dell PowerStore: Microsoft SQL Server Best Practices

 Management and monitoring

Management and monitoring

Introduction PowerStore offers multiple features for managing and monitoring storage.

Volume mapping After creating a volume, host mapping presents storage from the PowerStore array to a
and clustering host or a cluster of hosts that are defined in a host group. Follow general best practices as
outlined in the PowerStore Host Configuration Guide when mapping storage volumes.

Clustered SQL Server instances require special consideration since correct host mapping
is critical to cluster failover. SQL Server Failover Cluster Instances (FCI) rely on Windows
Server Failover Clustering (WSFC). Dell Technologies recommends using host groups to
uniformly present storage to all the initiators for reduced management complexity. This
allows a volume or set of volumes to be mapped to multiple hosts at a time and maintain
the same Logical Unit Number (LUN) across all hosts. If volumes in the failover cluster do
not have the same LUN number across Windows hosts in the cluster, failover does not
work properly. Once the cluster configuration is complete, test the cluster failover to
ensure that all failover components, including storage, are functioning as intended.

Creating PowerStore is virtualized to take advantage of all the drives in the appliance, and in
volumes addition there are many types of files that are part of a SQL Server instance. These types
of data often have different performance and snapshot requirements. For performance-
sensitive applications, Dell Technologies recommends creating at least five volumes for
an instance of SQL Server as shown in the following table.

Table 3. Volume provisioning recommendations

Number of
Typical
volumes
File type performance Typical snapshot requirements
per
requirements
instance

User DB data 1+ Lower performance Frequent snapshots, same

may be acceptable volume group as log volumes

User DB 1+ High performance Frequent snapshots, same

transaction log required volume group as data volumes

Data root directory 1 Lower performance Infrequent snapshots,

(includes system may be acceptable independent schedule
DBs)

Tempdb data and 1 High performance No snapshots

transaction log required

Native SQL Server 1 Lower performance Snapshots optional, independent

backup may be acceptable schedule

Memory-optimized 1+ High performance Frequent snapshots, same

filegroup (if used) required schedule as log volume

Dell PowerStore: Microsoft SQL Server Best Practices 27 2

 Management and monitoring

Associate databases that span multiple LUNs within a volume group. This helps ensure
that other features such as snapshots, replication, and thin clones will be configured
properly and maintain data consistency.

There is a balance between manageability and performance that is unique to each

environment. Generally, fewer volumes are easier to manage while more volumes enable
better performance and optimize for advanced features such as storage replication,
filegroup backups, and piecemeal restores.

Performance critical volumes can be added to a Watchlist for enhanced visibility. To do

this, select the volumes and under the More Actions menu, select Add to Watchlist.

Figure 10. Adding volumes to the Watchlist

Once the volumes are added to the Watchlist, they will appear on the main PowerStore
Dashboard. This provides some basic performance data as well as the ability to navigate
directly to them by clicking the volume name.

28 Dell PowerStore: Microsoft SQL Server Best Practices

 Management and monitoring

Figure 11. Adding volumes to watchlist

You can also add hosts to the Watchlist in the same manner. If you add a host that
already has volumes on the list, those volumes will be combined under the host.

Dell PowerStore: Microsoft SQL Server Best Practices 29 2

 Management and monitoring

Figure 12. Adding hosts to watchlist

Allocation unit When formatting volumes, choose a 64 KB allocation unit size. This aligns with the 64 KB
size database extent size, which is the allocation unit that is used by SQL Server. The
allocation unit size should be the same when formatting volumes on Windows or Linux
platforms.

Disk-format wait With trim/unmap enabled (default setting in Windows Server), significant wait time occurs
time when formatting a PowerStore volume that is more than 1 TB. The larger the volume is,
the longer the format wait time is, which is a common situation with external storage. This
case applies to volumes formatted as NTFS or ReFS.

To avoid a long format wait time when mapping and formatting a large volume,
temporarily disable trim/unmap. This setting is disabled using the fsutil command from
a command prompt with administrator privileges.

Figure 13 shows the commands to query the state and to disable trim/unmap for NTFS
and ReFS volumes on a host. A DisableDeleteNotify value of 1 means that trim/unmap
is disabled, and long format wait times are avoided when performing a quick format.

Changing the state of DisableDeleteNotify does not require a host reboot to take effect.

30 Dell PowerStore: Microsoft SQL Server Best Practices

 Management and monitoring

Figure 13. Use the fsutil command to query or change the state of trim/unmap

Once the volume is formatted, reenable trim/unmap so the host can take advantage of
deleted space reclamation for NTFS volumes.

For more background information about trim/unmap with PowerStore, see the document
Dell PowerStore: Microsoft Hyper-V Best Practices.

Thin clones Thin clone uses a pointer-based technology to create a read/write copy of a volume,
volume group, or file system. For data in a snapshot to be accessed, a thin clone must be
created. Because the thin clone volume shares data blocks with the parent, the physical
used space of the child volume is just the delta changes from after it was created. Thin
clones are advantageous in a SQL Server environment because a SQL Server database
can be duplicated for testing purposes, all while consuming less storage. For example, if a
SQL Server admin must clone a multi-terabyte database server for a developer to run
tests, the database can be isolated, tested, and only consume blocks that changed.

Within the PowerStore architecture, thin clones have several advantages for storage
administrators.
• The thin clone can have a different data protection policy from the parent volume.
• The parent volume can be deleted, and the thin clones become their own resource.
• Clone databases for testing monthly patches or development.
• It is treated as a regular storage resource and therefore supports many data
services such as protection policies, snapshot rules, or replication.
• Can be refreshed quickly to bring latest production changes to lower environment.
• Can be restored quickly to a baseline after testing.

Dell PowerStore: Microsoft SQL Server Best Practices 31 3

 Management and monitoring

When cloning databases for SQL Server, ensure all data and log volumes for the
database are on the same volume or are part of the same volume group. See best
practices for creating snapshots of SQL Server databases below. For more information,
see the document Dell PowerStore: Snapshots and Thin Clones.

Data encryption Many SQL Server applications have data encryption requirements, specifically on data at
rest. Data at Rest Encryption (D@RE) can be used as an encryption solution for SQL
Server without requiring any database or application changes. This also avoids any
potential performance impact to the database server or the applications and has no
performance impact on the array.

D@RE is enabled by default on the PowerStore array, so no configuration steps are

necessary to protect the drives.

Data reduction PowerStore includes zero detection, compression, and deduplication integrated in the
core product as part of the data reduction feature. While the amount of reduction varies
depending on the dataset, the overall savings are similar to or better than SQL Server
database compression. Since data reduction is always enabled on the array, SQL Server
database compression and backup compression can typically be disabled, saving CPU
resources on SQL Server.

When running SQL Server compute workloads on PowerStore X in AppsON mode, it is

recommended to disable SQL Server compression to allow CPU resources to be
allocated to other activities on the appliance.

For more information about the PowerStore data reduction benefits, see the document
PowerStore Data Efficiencies on Dell.com/StorageResources.

Scripting and Scripting and automation are often used to improve quality of repetitive tasks and
automation increase speed of deployment. The PowerStore platform supports several different tools
for administrators who want to automate management tasks or implementing
Infrastructure as Code (IaC).

REST API
The PowerStore REST API provides a complete interface for those wanting to perform
automation or custom applications for the PowerStore platform. Going to
https://<PowerStore Mgmt IP>/swaggerui in a browser provides an interactive API
reference for discovering available commands and viewing sample usage.

PowerStore CLI
The PowerStore CLI or PSTCLI is a command-line interface that is provided for managing
PowerStore systems. The PSTCLI provides a convenient way to manage PowerStore
without the complexity of using a programming interface. For more information, see the
document Dell PowerStore Command Line Reference at Dell.com/powerstoredocs.

Ansible
Ansible is a popular tool for automating IT infrastructure. Since Ansible is widely
supported by various hardware and software platforms, it makes it ideal for deployment

32 Dell PowerStore: Microsoft SQL Server Best Practices

 Data protection and disaster recovery

automation and IaC scenarios. The Ansible plug in for PowerStore as well as complete
documentation can be found at GitHub.com/Dell/Ansible-PowerStore.

Data protection and disaster recovery

Introduction PowerStore has integrated snapshot and replication capabilities to protect data, and is all
policy driven for ease of administration.

Snapshots and To automate and simplify protecting data, PowerStore uses protection policies. These
recoveries protection policies help to protect data, set retention policies, and help guarantee recovery
point objectives for an organization. Protection policies are a set of snapshot and
replication rules that are applied to a volume or group of volumes.

In addition, protection policies can be applied to volume groups. By applying a protection

policy to a volume group, it allows snapshots, replication, and recovery of the entire
group. This allows the protection of a database that spans multiple volumes. To ensure
successful recovery of SQL Server databases, ensure all database files, including the
transaction log file, are in the same volume group and the write-order consistency option
is enabled on the volume group.

Snapshots and Using array-based snapshots is an effective way to protect virtual machine data and
application establish a recovery point objective (RPO). In the PowerStore architecture, the snapshot
backup and schedule is created using protection policies. Each protection policy can define snapshot
restore rules to establish a schedule and retention, and replication rules to specify a destination
array and RPO.

Figure 14. Protection policies screen in PowerStore Manager

PowerStore has three data recovery mechanisms that all behave differently depending on
the usage scenario.
• Thin Clone: This takes an existing snapshot from a parent volume or volume
group and creates a child volume or volume group from that point in time.
• Refresh: Using the refresh operation, snapshot data can replace existing data
in the volume group. The existing data is removed, and snapshot data from the
new source is copied to it in-place. A parent volume or volume group can
refresh a child, and a child can refresh a parent.

Dell PowerStore: Microsoft SQL Server Best Practices 33 3

 Data protection and disaster recovery

• Restore: The restore operation replaces the contents of a parent storage

resource with data from an associated snapshot. Restoring resets the data in
the parent storage resource to the point in time the snapshot was taken.

Caution: Use of the refresh and restore operations on active database volumes may cause
unexpected results and behaviors. All host access to the volume must be ceased before
attempting these operations by either shutting down the SQL Server instance or taking the SQL
Server databases offline.

Figure 15. PowerStore snapshot and recovery anatomy

In addition to using snapshots to protect databases, if SQL Server hosts are leveraging
boot-from-SAN, the boot volume can be added to the same volume group as the
database volumes and the entire host can be protected.

Crash When taking array-based snapshots of virtual machines, it is important to remember that
consistency and snapshots that are taken without application coordination are considered crash consistent.
application Crash consistency is the storage term for data that has a snapshot taken in-flight without
consistency application awareness. While most modern applications can recover from crash-
consistent data, their recovery can yield varying levels of success. For example, when
recovering a Microsoft Windows virtual machine, as the operating system boots, it
behaves like it experienced unexpected power loss, and potentially invoke check disk
(chkdsk) on startup.

Dell AppSync enables taking application-consistent snapshots. When using products such
as AppSync, coordination between the array and the application can help to assure that
the data is quiesced, the caches are flushed, and the data is preserved in a known-good
state. Application consistent snapshots offer a higher probability of recovery success.

Replication and PowerStore offers asynchronous replication to transfer data to another PowerStore array.
remote recovery When the secondary array is at a different location, this can help to protect data from
localized geographical disasters. Replication RPOs and options are set within protection
policies.

34 Dell PowerStore: Microsoft SQL Server Best Practices

 Data protection and disaster recovery

Integrated copy AppSync is software that enables integrated Copy Data Management (iCDM) with the Dell
data primary storage systems, including PowerStore. AppSync, integrated with PowerStore
management snapshots, simplifies and automates the process of generating, consuming, and
managing copies of production data. This solution addresses copy management use
cases in a consolidated SQL database environment for database recovery and database
repurposing. AppSync automatically discovers application databases, learns the database
structure, and maps it through the virtualization layer to the underlying PowerStore
storage. Orchestration of all the activities required from copy creation and validation
through mounting the snapshots at the target host and launching or recovering the
application. This solution supports and simplifies Microsoft SQL Server workflows that
include refreshing, expiring, and restoring the production database. Additional information
about AppSync can be found in the documents AppSync Integration with Microsoft SQL
Server and AppSync Performance and Scalability Guidelines.

Dell PowerStore: Microsoft SQL Server Best Practices 35 3

 References

References
Related Dell.com/support is focused on meeting customer needs with proven services and
documentation support.

The following links provide additional related resources:

• PowerStore Info Hub
• Storage Class Memory in SQL Server 2016 SP1
• Dell NVDIMM-N Persistent Memory User Guide
• Dell SQL Server Solutions
• Microsoft SQL Server
• GitHub.com/Dell/Ansible-PowerStore
• AppSync Integration with Microsoft SQL Server
• AppSync Performance and Scalability Guidelines
• VMware Documentation
• SQL Server on VMware Best Practices Guide

36 Dell PowerStore: Microsoft SQL Server Best Practices