infortrend vmware Çözümleri – best practice

Copyright © 2016 Infortrend Technology, Inc. All rights reserved. Infortrend, ESVA, EonStor, EonNAS, and EonPath are trademarks or registered trademarks of Infortrend. All other marks and names mentioned herein may be trademarks of their respective owners. The information contained herein is subject to change without notice. Content provided as is, without express or implied warranties of any kind.

Infortrend VMware Solution – Best Practice

Version: 1.1

Updated: Mar, 2016

Abstract:

This document describes the best practices for using Infortrend storage in

a VMware environment. Infortrend storage provides a high degree of

compatibility and entire end-to-end virtual desk solution with VMware,

which not only centralizes the manageability but increases efficiency and

reliability.

Infortrend VMware Solutions

Best Practice


Copyright © 2016 Infortrend Technology, Inc. All rights reserved. / 71

Table of Contents

EonStor DS Storage Systems in VMware Virtualized Environments..................................... 3

Overview ............................................................................................................................... 3

EonStor DS Configuration Guideline for VMware ................................................................... 4

Configuration in SANWatch .................................................................................................. 4

Configuration in VMware vCenter ....................................................................................... 12

Configuration Considerations .............................................................................................. 14

Data formats ................................................................................................................ 14

Deployment of VMFS volumes .................................................................................... 18

RAID level .................................................................................................................... 19

VMware Native Multi-pathing Configuration ........................................................................ 20

Comprehensive integration between EonStor DS and VMware ........................................... 22

Hardware Acceleration - VMware vStorage APIs for Array Integration (VAAI) .................... 22

Introduction .................................................................................................................. 22

VAAI Primitives ............................................................................................................ 22

VSphere APIs for Storage Awareness (VASA) .................................................................... 27

Introduction .................................................................................................................. 27

Benefits ........................................................................................................................ 28

Requirements .............................................................................................................. 28

Installation ................................................................................................................... 29

Configuring VASA Provider .......................................................................................... 30

SSD Swap Cache ............................................................................................................... 36

Introduction .................................................................................................................. 36

Configuration ............................................................................................................... 36

Backup and Restore with vCenter Site Recovery Manager ................................................ 37

Introduction .................................................................................................................. 37

SRM and SRA Prerequisites ........................................................................................ 38

SRM and SRA Configuration ....................................................................................... 39

Workflows between SRM and EonStor DS .................................................................. 55

Reprotecting Virtual Machines after a Recovery .......................................................... 60

Best Practice and Recommendations on SRM ............................................................ 61

Remote Replication without SRM ....................................................................................... 62

Introduction .................................................................................................................. 62

Configuration ............................................................................................................... 64

Usage Case ................................................................................................................. 67

Difference between with and without SRM ......................................................................... 70

Scenario 1: Replicated Data Validation ........................................................................ 70

Scenario 2: After a Disaster ......................................................................................... 71

Conclusion ............................................................................................................................... 71



EonStor DS Storage Systems in VMware Virtualized

Environments

Overview

The increasing prevalence of virtualization is affecting businesses of all sizes and generating

new storage needs. Supporting large numbers of virtual machines and their valuable datasets in

VMware’s virtual infrastructure requires a fully-integrated SAN delivering high performance,

availability and agility. Infortrend EonStor DS integrates with key VMware vSphere features and

APIs to give users optimal resource efficiency, application productivity, operational resilience and

management simplicity.

To ensure the seamless integration of EonStor DS SAN storage into VMware virtualized

environments, Infortrend has performed comprehensive testing to verify its compatibility with

VMware products. For details about compatibility, please refer to the VMWare Compatibility

Guide or Infortrend’s EonStor DS Compatibility Matrix.

Effective software suites optimize and manage IT environments through virtualization. They

greatly reduce operating costs and increase IT service availability, security, and scalability, while

providing the flexibility to choose any OS, application, and hardware. VMware vSphere delivers

many enhanced features in terms of management, application services, infrastructure services,

compatibility, and third party extensibility. For more details about VMware products, please refer

to www.vmware.com.

http://www.vmware.com/resources/compatibility/search.php?rct=j&q=vmware%20compatibility&source=web&cd=1&ved=0CCkQFjAA&url=http://www.vmware.com/go/hcl&ei=66MLT6vvBMKHmQWP88j8BQ&usg=AFQjCNEXuH3H1BNILfm7AMoBNO0jQXhT-g

http://www.vmware.com/resources/compatibility/search.php?rct=j&q=vmware%20compatibility&source=web&cd=1&ved=0CCkQFjAA&url=http://www.vmware.com/go/hcl&ei=66MLT6vvBMKHmQWP88j8BQ&usg=AFQjCNEXuH3H1BNILfm7AMoBNO0jQXhT-g

http://www.vmware.com/



EonStor DS Configuration Guideline for VMware

Configuration in SANWatch

SANWatch, bundled with storage systems at no additional cost, is a browser-based

cross-platform software management suite that provides support for many mainstream operating

systems. It features SSL (Secure Socket Layer) link protection and role-based access levels for

management carried out on different installation sites and for different purposes.

Step 1: Launch Infortrend’s SANWatch on any server, configured on the same network subnet

as the targeted storage, to begin the partition creation for VMware datastore.

Step 2: Select the EonStor DS model (we use ESDS 3024 as example) and check its status.

Step 3: Select Logical Volumes and click Create Logical Volume.

Step 4: Create a name for the Logical Volume, assign to Slot A or B (Controller A or B) and

select targeted drive as the member drives using RAID protection (ex: RAID 1, 5, 6,

etc.…). Then click ‘Next’.

Note: SANWatch and User Guide download link can be found in http://support.infortrend.com/

Step 1 Step 2

http://support.infortrend.com/



Step 5: Click Create Partition under targeted Logical Volume to create a partition.

Step 6: Create the partition, set the partition name and the partition size based on requirement.

Then click ‘Next’.

Step 7: Click Host LUN Mapping to map partition as targeted partition, to the ESXi server.

Step 8: Select Customize the host LUN mapping configurations, Fibre or iSCSI interface, and

mapping channel or use extended LUN mapping. Then click OK.

Step 3 Step 4

Step 5 Step 6



Step 7 Step 8

Scenario 1

If you want to use storage as a space to store datastore or as a test environment, we

recommend using automatic mapping as following:



vCenter Server 5.5

ESXi 5.5

Network

DS 1016T

Management IP: 172.24.0.50

Host Channel IP:

CH4: 172.24.0.20 (Slot A)

CH4: 172.24.0.21 (Slot B)

IP: 172.24.0.14

IP: 172.24.0.3

1.45T

Step 9: Select Channels -> Click Host Channel Setting.

Scenario 2

If you want to configure LUM mapping with a higher security consideration and deploy related

VMware features, we recommend using Extended Host LUN mapping. We will introduce how

to configure EonStor DS 1016 to set-up 10 GbE iSCSI Extended LUN mapping and make

one 1.45 TB partition to be used as a datastore for one ESXi server below:

Configuration Reference

Note: The following setting can be skipped using EonStor DS with Fibre Channel interface.



Step 10: Click Configure in the targeted Channels.

Step 11: Configure IP address for each host board port.



Step 12: Go back to Host LUN mapping and select Use Extended Host LUN Functionality.

Step 13: The advanced LUN mapping provides array-based LUN access control over SAN for

host. The advanced LUN mapping binds a logical partition with a specific HBA port

and presents the partition to the HBA port. You have to add IQN of ESXi Host HBA

port after clicking Configure iSCSI Initiator Alias.



Note 1: How to find the IQN in vCenter Client:

Select targeted ESXi host -> Configuration tap -> Hardware -> Storage Adapters ->

Rescan all device -> Select targeted iSCSI Adapter Port -> Click Properties

Note 2: We only attach one cable to the HBA port in the following example, so there is only

one IQN shown.

On General page click Configure, there you can use copy and paste to get the IQN.



Step 14: After getting the IQN, go back to SANWatch to add the IQN: Press Add to add this IQN

and give it a specific Alias for recognition, then Click OK. Now you can see that the

IQN has been added.

You can check the status of mapping below the Partition Information.



Configuration in VMware vCenter

Step 1. Login to vCenter -> Home -> Hosts and Clusters -> Select Targeted ESXi host ->

Configuration tap -> Hardware -> Storage -> Rescan all device -> Add Storage

Step 2. Choose Storage Type Disk/LUN.

Step 3. Any unallocated storage devices will appear in the list. Select the required storage

device then click ‘Next’.

Step 2 Step 3



Step 4. Select the suitable file system version then click ‘Next’

You can refer to How VMFS5 Differs from VMFS3, for further information.

Step 5 Enter a name for this datastore then click ‘Next’.

Step 6. Set the capacity for this datastore then click ‘Next’.

Step 7. Click Refresh or Rescan All. Then you can see the datastore that you just created in

SANWatch. Now it can be used for this ESXi Server.

Step 4 Step 5

Step 6 Step 7

https://pubs.vmware.com/vsphere-50/index.jsp?topic=%2Fcom.vmware.vsphere.storage.doc_50%2FGUID-7552DAD4-1809-4687-B46E-ED9BB42CE277.html



Configuration Considerations

Data formats

Data volumes on EonStor DS Series arrays have to be configured as either a VMFS (Virtual

Machine File System) volume or RDM (Raw Device Mapping) volume 1 to make them

accessible to ESX Server.

VMFS is the VMware proprietary clustered file system and the most common access method. If

users would like to allow multiple virtual machines and multiple physical servers to access a

single volume, they should configure using the VMFS format.

Another way to allow virtual machines to access data volumes on storage is RDM. Virtual

machines access VMFS volumes and RDM volumes in different ways. As shown in the

following Figure, virtual machines can directly access a virtual disk in the VMFS format, but

their access to the RDM volume is enabled through a mapping file in the VMFS volume. This

mapping file contains metadata that redirects disk access to physical devices.

Accessing a VMFS volume and an RDM volume

Treating the RDM volume as a local disk, virtual machines can format it in the proper

way. RDM is especially useful in the following applications:

1. SAN-based snapshot/volume copy or other layered applications on virtual

machines.

2. Leveraging Microsoft Clustering Services (MSCS) to implement virtual-to-virtual

clusters or physical-to-virtual clusters. Clustered data and quorum disks have to be

configured as RDM volumes.

Followed by the step-by-step guide to create virtual machines with RDMs in the

vSphere Client.

1 Maximum size of an RDM volume out of 2TB



vCenter Server 5.5

ESXi 5.5

Network

DS 1016T

Management IP: 172.24.0.50

Host Channel IP:

CH4: 172.24.0.20 (Slot A)

CH4: 172.24.0.21 (Slot B)

IP: 172.24.0.14

IP: 172.24.0.3

1T

Vi r t u a l i z a t i o n l a ye r

OS

Apps

VM

OS

Apps

VM

OS

Apps

VM

OS

Apps

VM

Example to create Virtual Machines with RDMs

Step 1

After you finish configuring the storage and have added devices to vCenter, rescan the

HBAs on all your ESXi servers.

Step 2

Right-click on the targeted virtual machine, and click Edit Settings... to open the Virtual

Machine Properties window.



Step 3

Add a Hard Disk to a virtual machine and select Raw Device Mapping.

Step 4

Select the targeted RAW LUN and the datastore to store the LUN mapping.



Step 5

Select the compatibility mode based on your requirement:

Difference between Physical and Virtual compatibility RDMs

Now you can see the newly added RAW LUN.

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2009226



Deployment of VMFS volumes

The following guideline can help you properly deploying VMFS volumes:

1. Virtual machine boot disks and application data should be stored in separate VMFS

volumes. Most I/O issued to boot disks involves paging activities and is sensitive to

response time. By separating boot disks from application data, the risk of prolonged

response time due to application related I/O can be mitigated.

2. Database platforms for enterprise data management, such as Microsoft SQL

Server or Oracle, often use active logs and/or recovery data structures to track data

changes. In cases of unplanned application or operating system disruptions, these

active logs or recovery data structures are critical in ensuring system recovery and

data consistency. Therefore, all virtual machines supporting such database

platforms should be provided with an independent VMFS volume for storing active

log files and recovery data structures. Furthermore, if files or structures are

mirrored, the source and the target should be stored in separate VMFS volumes.

3. Application data, including database files, should be stored in a separate VMware

file system. This file system should not contain any structures that are critical for

application and/or database recovery.

4. It is recommended that VMFS volumes are no more than 80% full. This ensures

that administrators do not suddenly run out of space to accommodate user data

and VMware snapshots for virtual machines.



RAID level

EonStor DS Series storage arrays allow users to protect data volumes with various

RAID levels, including RAID 1, RAID 3, RAID 5, RAID 6, and RAID 10. Data volumes

in the same storage array can be protected with different RAID levels. The following

are general guidelines to configure RAID levels for data volumes in a VMware

virtualized environment:

1. Virtual machine boot volumes are generally subject to low I/O rates. They can be

configured with RAID 6 protection.

2. Large file servers with the vast majority of the storage consumed by static files can be

provided with RAID 6-protected volumes since I/O rates are expected to be low.

3. Infrastructure servers, such as Domain Name System (DNS), perform most of their

activities utilizing CPU and RAM and therefore are often subject to low I/O rates. If

users use virtual machines as infrastructure servers, it is proper to provide them

with RAID 6-protected volumes as storage space.

4. For most applications, RAID 5 is a proper level to protect virtual disks with.

However, if the application involves extensive logins, such as financial applications,

RAID 10 may be a better option as it works best for heavy transactional workloads

with high random writes.

5. Log devices for databases should be RAID 10-protected volumes. Furthermore, if

databases or application logs are mirrored, the source and the target should be

located on separate sets of disks (in VMFS format, if applicable).

6. Virtual machines that generate high workloads of small-block random read I/O,

such as Microsoft Exchange, should be allocated RAID 10-protected volumes for

better performance.



VMware Native Multi-pathing Configuration

Multi-pathing is a fault tolerant technique where there is more than one physical path between

the CPU in a computer system and its mass storage devices. Redundant physical path

components include controllers, switches, and the bridge devices connecting them. In addition to

fault tolerance, a multi-pathing solution redistributes the I/O request load among the multiple

paths between the host and the storage devices. This helps remove bottlenecks and balances

workloads.

By default, ESXi provides an extensible multi-pathing module called the Native Multi-pathing

Plug-In (NMP), which supports three types of Multi-path I/O, including Most Recently Used

(MRU), Fixed, and Round Robin.

Most Recently Used (MRU): Selects the first working path discovered upon the boot of

ESXi. If this active path becomes unavailable, the ESXi host will switch to another alternate

path and not return to the previous path even if it is fixed.

Fixed: Set a dedicated path to transfer I/O or NMP will select the first working path

discovered upon the boot of ESXi. If this active path becomes unavailable, the ESXi host will

switch to another available path and return to the previous path when it is recovered.

Round Robin (RR): Automatically selects all available paths and sends the I/O to each in a

circular fashion, which allows load balancing. All available paths are Active (I/O) and used

for transferring data.

Step 1:

Select Targeted ESXi host -> Configuration tap -> Hardware -> Storage Adapters -> Rescan all

device -> Right click targeted device and select Manage Paths



Step 2:

In Manage Paths, you can select the suitable multi-pathing solution based on your requirements.

Round Robin is highly recommended because all available paths can be used for transferring

data with load balancing.



Comprehensive integration between EonStor DS and

VMware

Hardware Acceleration - VMware vStorage APIs for Array Integration (VAAI)

Introduction

VAAI is a set of APIs (application programming interfaces) that allow the virtualization hypervisor

to offload I/O processing for certain data-related services to VAAI-supported storage systems,

such as the Infortrend EonStor DS family storage systems. It enables VMware ESX hosts to

convey virtual machine operations to storage hardware at a meta-level instead the traditional

data level reducing operational latency and traffic on a Fibre Channel or iSCSI storage network.

The topology of VAAI

VAAI Primitives

Full Copy - Speed up the virtual machine clone and migration tasks.

Block Zeroing - Reduce the host overhead associated with initialization job resulting from

space allocation in virtual machine creation.

Hardware Assisted Locking - Solve the virtual machine locking issue of competing for the

same storage resources.

Infortrend storage system works seamlessly with VMware’s ESX/ESXi virtualization hypervisor to

improve the efficiency of tasks such as Storage vMotion, cloning and provisioning new virtual

machines, all while reducing SCSI reservation locks and increasing the storage scalability by

adding new related SCSI commands to the standard SCSI command set. All enhancements

mentioned above have related primitives and SCSI commands for storage system integration.

The options for Infortrend storage to support VAAI are enabled by default. Users can check the

VAAI support in their storage system via vSphere Client or vSphere web client. There is a column



that shows the Hardware Acceleration status of each volume. If the storage has already enabled

VAAI support, the column will show “Supported”.



Full Copy

The primitive is also called hardware-accelerated copy, or copy offload, and it leverages the

standard SCSI XCOPY command. In most cases, deploying, cloning or migrating an existing

virtual machine with application data is extremely time-consuming. The primitive lets those

processes be entirely executed in the storage array, offloading the hypervisor server, resolving

the host overhead associated with VMware initiating and managing cloning or migrating

operations. It can significant reduce overall traffic on host and network.

The full copy primitive is helpful for tasks such as virtual machine data cloning, or VMDK file

migrating among datastores within a storage system using VMware Storage vMotion function.

The storage system will help to reduce the number of disk I/O requests from host - lowering the

host-side I/O bandwidth use during copy related operations.

In our Full copy test with EonStor DS 3000T, we cloned 4 virtual machines and 10 virtual

machines from one datastore to another. The more virtual machines we cloned, the more

significant the time saved (from 22% to 40.24%).

4 virtual machines cloned 10 virtual machines cloned

The difference of data copying or migration with or without VAAI support. It can significant reduce the I/O requests from host to storage array



Block zeroing

This primitive is also called hardware-accelerated zero, or zero offload, and it leverages the

standard SCSI WRITE-SAME command. When user tries to create a new virtual machine or new

virtual disks, a format called eager zeroed thick can be used for space pre-allocated and

pre-zeroed. The zeroed task is to write data with zero to clean and initialize space for new space

allocation. This means that it takes much longer to provision such virtual disks because it’s a

slow and inefficient process to zero out each disk space. The virtualization server must complete

all the zero writes of the entire virtual disk before it reports back that the disk zeroing is complete,

and it’s especially time-consuming for large disks. With the block zeroing primitive, these zeroing

operations can be offloaded to the storage system without issuing multiple zero writing

commands.

We test Block zeroing by creating 100GB virtual machines in “eager zeroed thick” virtual disk

format. With VAAI Block Zeroing primitive support, the virtual machine creation time was reduced

by over 20%.

he difference of block zeroing between without or with VAAI support.

VAAI helps to reduce lots of zeroing write commands issued by server.



Hardware-Assisted Locking

If the storage system was used for sharing to multiple virtualization servers, VMware’s file

system needs to coordinate all accesses from multiple servers. It allocates part of the storage

available to it for data describing virtual machines and their configurations, as well as the virtual

disks that they try to access. The implemented locking mechanism within the datastores that are

used to prevent any virtual machine from being run on, or modified by more than one host at a

time. Rather than locking the entire LUN, Hardware-Assisted Locking (ATS) provides a granular

LUN locking method and only locks the blocks on the LUN that are allocated to specific VMDK

used by a specific virtual machine. The initial implementation of mutual exclusion for updates to

these locking structures was built using SCSI RESERVE and RELEASE commands. Under the

protection of those locking methods, one host could update its data records on the device without

the risk of interference from other hosts attempting to update the same records. Multiple hosts

still can access its LUNs concurrently and also reduces the number of commands required to

successfully acquire on-disk locks.

In this test case, we deployed 100 virtual machines from 4 ESXi servers. All the servers share

the resources from one Infortrend storage system. We powered on all of the virtual machines

with a single shared datastore. Now we can see how it helps to reduce the bottleneck caused by

LUN locking.



VSphere APIs for Storage Awareness (VASA)

Introduction

One of the challenges facing VMware administrators is that it is time consuming to recognize the

capabilities and configurations of storage arrays. Infortrend VASA Provider is implemented to

allow VMware administrators to gain insight into the EonStor DS families directly from vCenter,

and assist them to efficiently manage storage resources.

This APIs provide a way of delivering storage arrays information to VMware vCenter, including:

Storage Capabilities

Collects and communicates information about physical capabilities and services offered by

the underlying storage system.

Storage Capabilities Profile



Difference: with and without VASA Provider

Storage Status

Reports on the status of various storage entities.

Benefits

VASA makes datastore management in virtual environment simple, by helping with:

Making decisions about the placement of datastore.

Making an early plan for space allocation.

Monitoring the status of storage in vCenter.

Requirements

vCenter Server version 5.5 or later

ESX/ESXi hosts version 5.5 or later

EonStor DS VASA Setup 1.0.6 or later

Suggest FW: 5.1x or later

The server that will be installed with VASA Provider must be installed with the following

components:

JAVA Runtime Environment 8 or later

Microsoft.NET Framework 2.5 or later



Installation

Step 1: Download Infortrend VASA Provider from Infortrend Customer Support System:


Step 2: Use an administrative account to log in to the Windows server (VMs or physical servers)

on which you are installing Infortrend VASA Provider.

Step 3: Locate the downloaded installer, and double-click it to start the installation process.

Follow the onscreen instructions while clicking ‘Next’ to proceed.

Step 4: When you are prompted to select the install location, use the default location or click

Browse to select the installation location, and then click ‘Next’.

Step 3 Step 4

Note: VASA Provider and vCenter Server cannot be installed on the same server.




Configuring VASA Provider

VASA Provider must be configured to allow vCenter Server to communicate with the storage

subsystems and monitor their status.

Step 1: Adding or removing storage subsystems

1. Double-click IFT VASA Raid IP Manager on your Windows desktop to start the application.

2. To add storage subsystems and edit the user credentials for accessing VASA Provider

double-click a cell in the Set RAID IP tab window and enter the IP address of the storage

subsystem that will be managed by vCenter Server. If there is more than one storage

subsystem in your storage network, double-click an empty cell and repeat the same process.

Click ‘Save’ when finished.

Setup the sync interval (in minutes) from the drop-down menu to let VASA provider collect

the most updated information from the storage subsystem on a regular basis.



3. On the General Setting tab, enter a user-defined Username and Password, or use the

default user credentials for accessing VASA Provider: "username" for the Username field

and "password" for the Password field.

Note: JAVA Runtime Environment 8 or later is required to be installed before VASA

installation, and you have to confirm the correct JRE install path. In addition, if JRE

install path has been modified, you have to go to $VASA_install_PATH/tomcat and

run uninstall.bat, then go back to IFT VASA Manager to re-install the VASA service.



Step 2: Registering Infortrend VASA Provider with vCenter Server

1. From the homepage of vSphere Client, click Storage Providers, click Add, enter the following

information, and then click ‘OK’.

Note:

Name: Name your VASA provider

URL: Enter the URL of your VASA provider.

For HTTPS (SSL) connection:

https://[vasa server IP address]:8443/vasa/services/vasaService

For HTTP connection:

http://[vasa server IP address]:8080/vasa/services/vasaService

Login and Password:

Enter the user credentials for accessing VASA Provider



2. From the homepage of vSphere Client, click Datastores and Datastore Clusters, select the

Summary tab, and check the information in the Storage Capabilities and Storage Capability

Details sections.

3. Go to homepage of vSphere Client -> Click VM Storage Profiles -> click Create VM Storage

Profile, and then do the following:

a. Name the profile and click ‘Next’.

b. Select the options to define the capabilities for the storage, and keep clicking ‘Next’ to

finish the setup.



4. From the homepage of vSphere Client, click VMs and Templates, right-click and select New

Virtual Machine from the contextual menu. In the window that appears, do the following:

a. Select Typical and click ‘Next’.

b. Choose the name of the profile you just created from the VM Storage.

c. Profile drop-down menu, and click ‘OK’.



5. To change the profile of a VM, right-click on it, choose VM Storage Profile > Manage Profiles

from the contextual menu, and then choose a profile form the drop-down menu. If you click

Propagate to disks, the storage profiles will be applied to all VMs stored on the same virtual

disk. Click OK to complete the setup.

6. From the homepage of vSphere Client, select VM Storage Profiles, and from the left sidebar,

select the profile you just created, and the list of all VMs applied with that profile will appear

under the Virtual Machines tab. If the profile does not match the actual capabilities of the

virtual machine, you will see "Noncompliant" under the Compliance Status column.



SSD Swap Cache

Introduction

vSphere 5.0 (or later) enables users to choose to configure a swap cache on the SSD. ESXi

server will use this swap cache to store the swapped-out pages instead of sending them to the

regular and slower hypervisor swap file on the disk. This feature in ESXi uses a relatively small

amount of solid-state drive (SSD) located in EonStor DS systems to significantly reduce the

performance impact of host-level memory swapping. In general, the host cache is made of files

equally divided into 1 GB size (.vswp). ESXi uses these files as write back cache for virtual

machine swap files.

Configuration

Step 1: Create LDs and partitions by using SSDs on EonStor DS, and mapping them to specific

ESXi hosts.

Step 2: Configure host cache for targeted ESXi.

Note: In order to enable Swap cache, add a custom value for hot cache allocation.



Backup and Restore with vCenter Site Recovery Manager

Introduction

One of the most complicated procedures of a Disaster Recovery (DR) Failover in a VMware

environment is the process of setting up inventory, register, reconfigure and power up VMs at the

DR site. VMware Site Recovery Manager is an ideal solution that simplifies and automates

disaster recovery. SRM is an easy-to-use graphical user interface (GUI) application designed to

manage virtual machines in datacenter and set recovery plan and policy in advance. The user

can, for example, define each VMs into High, Medium or Low priority groups and determine the

recovery procedure running user-defined scripts and start a sequence of individual VMs in

advance. When planned migration or unexpected disaster occurs, the system administrator can

easily initiate the recovery via VMware vSphere Web Client with the click of a button.

SRM enables integration with array-based replication, as well as the use of a native VMware

vSphere-based replication. SRM, a plug-in for vCenter on both the primary (or protected) site

and the recovery site, enables separate VMware environments to communicate and negotiate

with each other. In array-based replication, VMware SRM perfectly coordinates with EonStor DS

Remote Replication technology via Infortrend Storage Replication Adapter (SRA). Remote

replication capability allows users to create full data copies of virtual machines across storage

systems, in synchronous or asynchronous mode.

VMware SRM architecture



SRM and Storage Replication Adapter (SRA) should be installed both at the protected and

recovery site. Once SRA has been installed, the user should easily discover Infortrend SRA in

Array Managers of SRM and execute additional configurations for the local array at the protected

site and peer arrays at the recovery site. The following is the step-by-step guide of how to

configure SRM.

SRM and SRA Prerequisites

EonStor DS series requirements for SRA 1.0.8(i) and above

Remote replication license required at both the protected and recovery sites

Recommend systems with firmware 5.1x or later

VMware vCenter SRM requirements

VMware vCenter Site Recovery Manager 5.5 or 5.8 license

http://www.vmware.com/products/site-recovery-manager/pricing.html

Site Recovery Manager, VMware vCenter™ servers, and SRM database servers

should be installed in individual VMs or physical servers

SRA (installed on the server with SRM)

VMware tools for Guest OS virtual machine

Windows platform support only

VMware vCenter, vSphere Client, ESXi Server, Guest OS, Database are required,

Please check compatibility Matrixes for SRM

https://www.vmware.com/support/srm/srm-compat-matrix-5-0.html

http://www.vmware.com/products/site-recovery-manager/pricing.html




SRM and SRA Configuration

vCenter Server 5.5

ESXi 5.5 ESXi 5.5

vCenter Server 5.5

DS 3016T DS 1016T

ESXi IP:

172.24.0.13

vCenter IP:

172.24.0.3

DS 3016T IP:

172.24.0.100

ESXi IP:

172.24.0.14

vCenter IP:

172.24.12.73

Protected Site

DS 1016T IP:

172.24.0.50

Protected Site Recovery Site

Environment to demonstrate SRM

Step 1: Set up Replication Pair and create backup schedule.

1. Before SRM and SRA installation, you have to set up Replication Pair via SANWatch.

Connect two systems by in-band or source system with in-band and targeted system with

out-band and then select Replication Manager.

2. Create a replication pair.

3. Select a method for pair creation.



Scenario 1:

If you want to create a new source volume, select “Create a new volume as the source of

replication pair”.

Scenario 2:

If you want to use an existing source volume, choose “Select existing volume for replication

pair creation”.



4. Select your source device and source partition.

5. Select your remote device and logical volume on that remote device.

6. Select a replication mode

There are several replication types, including synchronous mirror, asynchronous mirror, and

volume copy. Synchronous mirror requires higher network bandwidth and is suitable for

critical data. Asynchronous mirror is suitable for archived data. Volume copy is used for

application development or tests. In most scenarios, we recommend using volume mirror in

asynchronous mode because of lower performance impact during office hours. The backup

schedule should be arranged based on user requirements.



7. Select Diagnose Network to check the connection status between the source device and

target device before creating the replication pair.

8. Click ‘Next’ and system will create a replication pair.

9. Create a backup schedule and select the backup type.

10. Select the schedule target and create the schedule.



Step 2. SRM Server Installation

1. Download Infortrend SRA from VMware or Infortrend Customer Support System

VMware website:

https://my.vmware.com/web/vmware/details?productId=357&downloadGroup=SRM_SRA

Infortrend Customer Support System:


2. Install SRM Server in VMs or physical servers on both Protected and Recovery Sites

In this best practice, we install SRM Server in a virtual machine (take ESXi 172.24.0.13 for

example)

vCenter Server 5.5

ESXi 5.5 ESXi 5.5

vCenter Server 5.5

DS 3016T DS 1016T

ESXi IP:

172.24.0.13

vCenter IP:

172.24.0.3

DS 3016T IP:

172.24.0.100

ESXi IP:

172.24.0.14

vCenter IP:

172.24.12.73

DS 1016T IP:

172.24.0.50


Virtualization layer Virtualization layer

OS

Apps

VM

OS

Apps

VM

SRM Server Protected VM

OS

Apps

VM

SRM Server

Before starting SRM server installation, we have to set up a dedicated database for SRM server

(ESXi 172.24.0.13). Please refer to the VMware website for database compatibility matrixes.



https://my.vmware.com/web/vmware/details?productId=357&downloadGroup=SRM_SRA






Open the Guest OS ESXi (172.24.0.13) console to install SRM, and setup vCenter Server

Credentials and database configuration



If SRM service failed to start during installation, please change the VMware Site Recovery

Manager Service in Windows Service Manager to use the domain account.



Step 3. Storage Replication Adapters 5.5/ 5.8 Installation

1. Download SRA from support site and install.

2. Please confirm the Infortrend DS Series SRA Service is started in SRM server.



Step 4. Enable SRM in vCenter

Go to Home -> Plug-ins -> Manage Plug-ins and download available SRM Plug-ins

Step 5. Click Site Recovery in home page and Configure Connection between Protected site

(172.24.0.3) and Recovery Site (172.24.12.73)

Go to Site -> Configure Connection -> Enter IP Address and access port number ->

Provide vCenter user name and password

Note: Please confirm port 80 is available and not used by other services in order to build

connection among these two sites.



Step 6. Right Click the Recovery Site’s IP to login.

Step 7. After connection creation, you must configure resource, folder, and network mapping to

make SRM move all protected data to the right place in the recovery site.

Go to Site -> Resource Mapping / Folder Mapping / Network Mapping

Note: Please confirm all the information including resource, folder and networking are

synchronized on both sites.



Step 8. Configure Placeholder Datastores

For every virtual machine in a protection group, Site Recovery Manager creates a

placeholder virtual machine at the recovery site. You must identify a datastore on the

recovery site in which Site Recovery Manager can store the placeholder virtual

machines.

Note: You cannot select replicated datastores as placeholder datastores. For clusters, the

placeholder datastores must be visible to all of the hosts in the cluster.



Step 9. Configure Array Manager

After establishing a basic connection between primary and remote sites, the user

should discover Infortrend SRA in Array Managers. Users may configure Local and

Peer array IP addresses.

Go to Array Managers -> Add an Array Manager -> Enter display name for this adapter,

such as Infortrend_SRA -> Enter both local and peer array’s IP address -> Repeat the

same procedure to add Array Manager on recovery site.



Note. Please enable Array Pairs manually on Discovered Array Pairs page as following:



Step 10. Create a Protection Group

Once both sites and Array Managers are configured, the user can create a

protection Group and select targeted partition(s) to protect.

Go to Protection Groups -> Create a Protection Group -> Select targeted array pair

-> Select targeted partition(s) to protect -> Naming this protection group



Step 11. Create Recovery Plan

Go to Recovery Plans -> Create Recovery Plan -> Select Recovery Site -> Select

targeted protection group -> SRM will automatically test network -> Naming this

recovery plan



Step 12.

Different interaction workflows between SRM server and EonStor DS series storage can be

initiated now, including Test, Cleanup, Recovery, and Reprotect. In addition, the user can define

the priority for each virtual machine. Site Recovery Manager sets all virtual machines in a new

recovery plan to recovery priority level 3 by default. If a planned migration or a disaster occurs,

Site Recovery Manager will start virtual machines on the recovery site based on the priority. Site

Recovery Manager starts priority 1 virtual machines first, then priority 2 virtual machines second,

and so on.

Note: SRA Parameters Configuration

Go to “Home > Solutions and Applications > Site

Recovery”, right click both protected site and recovery site, click “Advanced

settings”, go to “storage”, and change the default value 300 to 600 for the

parameter storage.commandTimeout.



Workflows between SRM and EonStor DS

Test Recovery Plan

When you test a recovery plan, Site Recovery Manager will execute the recovery plan step by

step, including test network and execute remote mirror of replicated data at the recovery site. It

will make sure the synchronized data is available and writeable at the recovery site. The most

improtant thing is that Site Recovery Manager does not disrupt the current operations at the

protected site. It will skip the steps in the plan to shutdown all virtual machines in protected site

and force devices at the recovery site to assume ownership.

Figure: Testing the recovery plan



Figure: Test Confirmation

Notes: Before the test starts you will get a warning to make sure you know what you are

doing. You have the option here to select a checkbox to synchronize the datastores

too.



Cleanup after Testing a Recovery Plan

After testing a recovery plan, running a cleanup operation will make SRA remove the test

environment and return the recovery plan back to Ready state.

Figure: Cleanup after testing

One-Click Recovery

There are two failover possibilities that may occur and will require SRM and SRA to act. One is

planned downtime, such as upgrades of operating system / applications and hardware

maintenance. In this instance, we would ask SRA to shut down the VMs, unmount the datastores

at protected site, execute datastores replication from source to target storage, and restart the

services at the recovery site. Another one is unplanned downtime when an unexpected event

occurs. In this scenario, minimizing the impact is the key. Once virtual machines are offline at the

protected site, copies of these virtual machines at the recovery site based on the last storage

replication from protected sites can be powered up to resume the services. By using one-click

button, both scenarios will be automatically covered by implementing the pre-defined procedure

to migrate or recover the protected data.

Figure: One-Click to execute planned migration or disaster recovery



For example, site A is the protected site and site B is the recovery site. If site A goes offline

because of disaster. Using one click of Disaster Recovery, Site Recovery Manager will resume

the services at site B based on the last storage replication.

Figure: Scenario of disaster recovery

SRM 5.8

1. From the home page of vSphere Client, click the Site Recovery icon to launch SRM. Select

Sites from the left sidebar.

2. Select the site to which you want to adjust the settings, and then select Manage >

Advanced Settings tab.

3. Select Recovery, and select the checkbox for "recovery.forceRecovery".

4. Repeat the above procedures for the opposite site.

SRM 5.5

1. From the home page of vSphere Client, click the Site Recovery icon to launch SRM.

2. Click Sites, right-click on the site to which you want to adjust the settings, and then choose

Advanced Settings.

3. Select Recovery, and select the checkbox for "recovery.forceRecovery".

4. Repeat the above procedures for the opposite site.

Note. Enabling forced recovery

If a disastrous event occurs and the protected site goes offline for an extended period of time,

making SRM unable to perform its standard tasks covered in the previous chapter, you can

follow the steps introduced in this chapter to carry out the recovery plan with a forced

recovery option. This chapter assumes you are already familiar with the standard tasks of a

recovery plan. For detailed instructions on the steps mentioned in this chapter, refer to the

previous chapters. For more information about forced recovery, see VMware's

documentation:

http://pubs.vmware.com/srm-58/index.jsp?topic=%2Fcom.vmware.srm.admin.doc%2FGUID-

5B7CDD0A-75D7-426D-AC62-FAB73DED78AD.html&resultof=%22forced%22%20%22forc

%22%20%22recovery%22%20

http://pubs.vmware.com/srm-58/index.jsp?topic=%2Fcom.vmware.srm.admin.doc%2FGUID-5B7CDD0A-75D7-426D-AC62-FAB73DED78AD.html&resultof=%22forced%22%20%22forc%22%20%22recovery%22%20





Reprotecting Virtual Machines after a Recovery

When site A comes back online, you can execute a Reprotect operation to protect the virtual

machines on site B. Site B becomes the protected site, and site A becomes the recovery

site. Site Recovery Manager and SRA reverse the direction of replication from site B to site A.

Figure: Reprotect after planned migration or disaster recovery

Figure: How it works in reprotect



Best Practice and Recommendations on SRM

1. Prepare an organized recovery plan in advance

i. Define the backup schedule to meet RTO and RPO requirements.

ii. Test the recovery plan to make sure it is usable.

iii. Define the importance of each virtual machine and assign each VMs into High,

Medium or Low priority groups and determine the start sequence of individual VMs

in advance.

2. SRA only supports replication among DS models.

3. Within the same data center or non-cross WAN environment, EonStor DS can execute

synchronous replication for business-critical service.

4. The user may choose asynchronous mode to allow the mirroring during off-peak hours in

order to achieve effective bandwidth utilization and to reduce costs.

5. To ensure successful remote replication in an array-based scenario, users are advised to

reserve extra space in the storage system at recovery sites.

6. It is highly recommend to configure Bi-directional protection. SRM is configured to fail over

virtual machines from site A to site B at the same time that it is configured to fail over a

different set of virtual machines from site B to site A.

7. For more information about Infortrend Remote Replication, please refer to

SANWatch Configurations for Remote Replication on EonStor DS

www.infortrend.com/ImageLoader/LoadDoc/189

EonStor DS remote replication feature guide

www.infortrend.com/ImageLoader/LoadDoc/598

Note:

Uni-directional protection to fail over virtual machines from site A to the site B. In this case,

licenses are required only for the protected virtual machines at protected site A. However,

SRM licenses must be purchased for the protected virtual machines at both sites for

Bi-directional protection.

http://www.infortrend.com/ImageLoader/LoadDoc/189

http://www.infortrend.com/ImageLoader/LoadDoc/598



Remote Replication without SRM

Introduction

Even without SRM, remote replication on EonStor DS storage systems offers strong defense

against major disruption to IT continuity, especially in the case of natural disasters. In

combination with data services, such as snapshot and volume copy, remote replication allows

entire data sets to be saved at physically separate locations. If one location becomes

inaccessible, data can still be used from the remotely replicated data set. Additionally, Infortrend

remote replication supports multiple modes and scenarios, with up to 64 replication pairs per

system for maximum flexibility and redundancy. Infortrend systems also feature shorter

replication times and higher bandwidth utilization to minimize delays and latency.

EonStor DS Remote Replication to protect virtual machines



OS

Apps

VM

Windows

Server

vCenter Server 5.5

ESXi 5.5 ESXi 5.5

vCenter Server 5.5

DS 3016T DS 1016T

ESXi IP:

172.24.0.13

vCenter IP:

172.24.0.3

DS 3016T IP:

172.24.0.100

ESXi IP:

172.24.0.14

vCenter IP:

172.24.12.73

DS 1016T IP:

172.24.0.50



OS

Apps

VM

Windows

Server

OS

Apps

VM

SUSE

Server

Datastore for Windows 1


Datastore for SUSE 1

RemoteMirror 1

RemoteMirror 2

RemoteMirror 3

Example to protect virtual machines by using Remote Replication



Configuration

Step 1: Set up a Replication Pair between protected site and recovery site, and create a backup

schedule via SANWatch as mentioned before.

Step 2: Replicated data validation

Currently we have set a pair between protected and recovery site. If you want to validate

replicated data, you have to take a snapshot of the LUN (take Partition 1 as example, which is

the target of Remote Mirror 1) on recovery site and map it to ESXi host (172.24.0.14) on recovery

site to validate the replicated data.

OS

Apps

VM

Windows

Server

vCenter Server 5.5

ESXi 5.5 ESXi 5.5

vCenter Server 5.5

DS 3016T DS 1016T

ESXi IP:

172.24.0.13

vCenter IP:

172.24.0.3

DS 3016T IP:

172.24.0.100

ESXi IP:

172.24.0.14

vCenter IP:

172.24.12.73

DS 1016T IP:

172.24.0.50



OS

Apps

VM

Windows

Server

OS

Apps

VM

SUSE

Server



Datastore for SUSE 1

RemoteMirror 1

RemoteMirror 2

RemoteMirror 3

Example to validate replicated data by using snapshot



Go to SANWatch -> Select the partition which is the target of Remote Replication -> Select

Snapshot -> Click Take Snapshot to take a snapshot for this partition -> When the snapshot has

been taken, select targeted snapshot and click Map -> Create host LUN mapping

Go back to vCenter -> Select targeted ESX host-> Configuration -> Storage -> Add Storage ->

Select VMFS Mount option (Keep the existing signature / assign a new signature) -> Select

Custom configuration



If you do not need to resign a VMFS datastore copy, you can mount it without changing its

signature. You can keep the signature if, for example, you maintain synchronized copies of

virtual machines at a secondary site as part of a disaster recovery plan. In the event of a disaster

at the primary site, you can mount the datastore copy and power on the virtual machines at the

secondary site.



Usage Case

Scenario 1: When a disaster occurs

Map the target LUNs to the remote ESX host manually, so that the service can continue.

Scenario 2: When local site has been recovered

Step 1: Unmap source when local has been recovered

Step 2: Switch source and target roles, and Sync from remote to local



Step 3: Stop service and unmap remote source

Step 4: Split

Step 5: Switch source and target roles



Step 6: Sync from local to remote

Step 7: Map source and service to complete failback procedure



Difference between with and without SRM

Scenario 1: Replicated Data Validation

It requires a complicated procedure to validate replicated data without SRM

1. Execute Remote Replication from Protected site to Recovery site.

2. Create Snapshot of the Target Volume.

3. Map the snapshot to the host.

4. Boot the host to check data validity.

5. Check data validity.

1. You only need to set remote replication pairs via SANWatch once in advance.

2. Click on the Test button and SRM will automatically check data validity.



Scenario 2: After a Disaster

It requires complicated procedures to resume services via multiple interfaces.

1. Based on the previous procedure of remote replication and mapping target volumes to ESXi

Host via Infortrend GUI.

2. Manually boot each virtual machine to resume service in vCenter.

However, SRM and EonStor DS Remote Replication make operations simple.

1. Based on previous procedure of remote replication, via SRM GUI, click on the ‘Recovery’

button and SRM will automatically launch the recovery process.

Conclusion

The integration of VMware with EonStor DS series delivers clear benefits that significantly

reduce the complexity of and workload in a virtual environment. For virtualized deployments of all

sizes, the combination of VMware and Infortrend solutions can greatly assist the IT administrator

to quickly deliver flexibility, performance, reliability and agility to the IT infrastructure in the most

cost-effective way. Therefore, the integration of VMware with Infortrend EonStor DS series is an

ideal virtualization solution.

infortrend vmware Çözümleri – best practice

Business