Storage Data Protection Technologies and

Applications
 Foreword

⚫ Traditional data protection solutions focus on periodic data backup. Therefore,

problems such as no backup window, inconsistent data, and impact on the
production system always occur.
⚫ This course describes storage data protection technologies such as HyperSnap,
HyperClone, HyperReplication, and HyperMetro which are new data protection
methods.

1 Huawei Confidential
 Objectives

On completion of this course, you will be able to understand the principles,

configuration methods, and application scenarios of the following features:
 HyperSnap
 HyperClone
 HyperReplication
 HyperMetro

2 Huawei Confidential
 Contents

1. HyperSnap

2. HyperClone

3. HyperReplication

4. HyperMetro

3 Huawei Confidential
 Overview

⚫ Storage Networking Industry Association (SNIA) defines a snapshot as

follows:
A snapshot is an available copy of the specified data collection. The copy
contains the image for the relevant data at a time point when the copy
begins.

A snapshot can be a duplicate or replicate of data.

4 Huawei Confidential

• Purposes of a snapshot:

▫ Backup and archiving: A snapshot can be used to serve as a data source for backup
and archiving.

▫ Quick recovery: A snapshot flexibly and frequently generates recovery points in time
for data on storage devices, enabling fast data recovery when necessary.

▫ Instant generation: A snapshot is instantaneously generated without interrupting

host services. It is a data duplicate of the source LUN at a specific point in time.
 Working Principles of HyperSnap
⚫ Definition: A snapshot is a consistent copy of the source data at a certain point in time. After the snapshot is generated, it
can be read by hosts and used as a data backup at a certain point in time.
⚫ Main features
 Instant generation: A storage system can generate a snapshot within a few seconds to obtain the consistent copy of source data.

 Small storage space occupation: A snapshot is not a full physical data copy, which does not occupy large storage space. Therefore, a
snapshot for a large amount of source data occupies only a small space.

a b c a b c
d e f d e f
8:00 AM
g h i g h i
j k l j k l

a b c a b c
d m f d e f 9:00 PM
g h n g h i
j k l j k l

5 Huawei Confidential

• Common snapshot terms:

▫ Source volume: A volume that stores the source data of a snapshot. It is presented as
a LUN to users.

▫ Snapshot volume: A data copy logically generated after a virtual snapshot is created
for a source LUN. It is presented as a LUN to users.

▫ Redirect on write: When data is modified, new space is allocated to new data. After
the new data has been written successfully, the original space is released.

▫ Snapshot rollback: Data of a snapshot LUN is copied to the source LUN. In this way,
data of the source LUN is recovered to state at the point in time when the snapshot
LUN was activated.

▫ Inactive: Status of a snapshot in which the snapshot is unavailable. The opposite

status is activated.
HyperSnap Principles – Zero Performance Loss
Data requested to be written to L2 of the
source LUN is written to P5. Data requested to be written Data requested to be written to L2
Data requested to be written to L2 of the to L0 of snapshot 1 is written of snapshot 2 is written to P8.
source LUN is again written to P7. to P6. ⚫ Data requested to be written to L2 of
the source LUN is written to a new
LUN Mapping Table Snap Mapping Table
space P5. The original space P2 is
Snap Mapping Table
referenced by the snapshot.
⚫ Data requested to be written to L0 of
snapshot 1 is written to the new
space P6, bringing no additional read
and write overhead.
⚫ When data is written to L2 of the
source LUN again, the requested data
is written to a new space P7. The
original space P5 is released because
L3->P3 L4->P4 L2->P5 L2->P7 L2->P8
L0->P6 L0->P0 L1->P1 L2->P2 it is not referenced by a snapshot.
⚫ A new snapshot 2 is created and
activated.
A B C D E F G H I

P0 P1 P2 P3 P4 P5 P6 P7 P8
SSD storage space

6 Huawei Confidential
HyperSnap Principles – Rollback
Data 10:00 AM Data 11:00 AM Virus Infection Data 11:00 AM
Restore

10 11 12 TIME

Snap 10:00 AM Snap 11:00 AM

7 Huawei Confidential
HyperSnap Principles – Snapshot Cascading and Cross-Level
Rollback
Source
⚫ Snapshot cascading: It is a child snapshot of
volume
a parent snapshot. The difference between
snapshot duplicates and snapshot cascading
is that the latter includes the data of its
parent snapshot. Other functions are the
8:00 9:00
same as common snapshots.
snapshot0 snapshot1
10:00 11:00
⚫ Cross-level rollback: Snapshots sharing the
Snapshot1.snapshot0 Snapshot1.snapshot1
same source volume can roll back each other
regardless of their cascading levels.

8 Huawei Confidential
Key Technologies of HyperSnap-Duplicate

How can I obtain multiple

Source duplicates of the same snapshot?
volume Snapshot

8:00

Snapshots are virtual, so they

8:00 8:00 8:00
can be duplicated fast.

9 Huawei Confidential
Key Technologies of HyperSnap - Rollback Before Write

How can I instantly recover data?

Source
volume Snapshot
Snapshot
rollback During the rollback, when a host
8:00
writes data to a source LUN, the
snapshot copies the data blocks to
the source LUN, and then the host
continues to write data.
When no host reads or writes data,
the snapshot data is rolled back to
the source volume in sequence.

10 Huawei Confidential
Application Scenario
Source Snapshot
LUN LUN
Source
LUN

Snapshot
Duplicate Duplicate Duplicate Duplicate
LUN

1:00
2:00
4:00 3:00

Report Data test Data Decision-

generation analysis making
support

⚫ Continuous data protection ⚫ Data backup and restoration

11 Huawei Confidential
 Configuration Process
Required
Start
Optional

1. Checking the
availability of the Check the license file.
snapshot function

Performing this operation when a

Create a source LUN. storage system does not have the
source LUN of a snapshot.

2. Creating a snapshot Create a snapshot.

End

12 Huawei Confidential

• Snapshot configuration terms:

▫ Activated: Status of a snapshot. This status indicates that the snapshot is available.

▫ Inactive: Status of a snapshot. This status indicates that the snapshot is unavailable.

▫ Reactivates a snapshot: Combination of the command for deactivating and activating

a snapshot.

▫ Snapshot consistency group: To ensure that all snapshots, simultaneously created by

multiple LUNs for the same type of services, are consistent in time, you need to add
these LUNs to a protection group (PG) and create a snapshot consistency group (CG)
for the PG for unified management.

• For details, see Huawei Data Storage Infocenter: http://support-

it.huawei.com/storage/#/home
 Contents

1. HyperSnap

2. HyperClone

3. HyperReplication

4. HyperMetro

13 Huawei Confidential
Overview
⚫ Definition

 HyperClone creates a full data copy (a target LUN) of a source LUN at a specified point in time
(synchronization start time).

⚫ Features

 A target LUN can be read and written during synchronization.

 Full synchronization and incremental synchronization are supported.

 Forward synchronization and reverse synchronization are supported.

 Consistency groups are supported.

14 Huawei Confidential
Working Principles of HyperClone
⚫ Definition: Clone is a consistent data copy of a source data at a specific point in time. It functions as a complete data copy
after data synchronization. It serves as a data backup and is accessible to hosts.
⚫ Main features:
 Quick clone generation: A storage system can generate a clone within several seconds to obtain a consistency copy of a source data. The
generated clone can be read and written immediately. Users can configure different deduplication and compression attributes for the
generated clone.

 Online splitting: A split can be performed to cancel the association between a source LUN and a clone LUN without interrupting services.
The split read and write operation on the clone LUN will not affect the I/O process of the source LUN.

Create Synchronize
HyperClone. HyperClone.
a b c a b c a b c
d e f d e f d e f
g h i g h i g h i
j k l j k l j k l

15 Huawei Confidential
HyperClone Principles - Synchronization
Scenario 1: Initial Scenario 2: Synchronization is
synchronization and full copy performed again after the first
are performed. synchronization, and differential copy
is performed.
a
b a a b Stored data
b
2 3
Snapshot c
1 2 a b c Newly copied data

Snapshot
1 c New data written to the host
a a
b b a a Internal signal flow of the storage
b b system
Source Target
LUN LUN c c
HyperClone pair
1 Create a HyperClone pair. Source Target
LUN LUN
2 Create a snapshot for the source LUN
after synchronization is started. 1 Create a snapshot for the source LUN after
a second synchronization.
3 Copy all data a and b to the target LUN. 2 Copy incremental data c to the target LUN.

16 Huawei Confidential
 HyperClone Principles - Reverse Synchronization
Scenario 1: Full copy Scenario 2: Differential copy a b c Stored data

a a a b c d Newly copied data

b b
c c
d New data written to the host
d d
2 1 2 1 Internal signal flow of the
Snapshot Snapshot storage system

HyperClone pair

a a a a
b b b b
c c c c
d d d d
Source Target LUN Source Target LUN
LUN LUN
1 Create a snapshot for the target LUN after the 1 Create a snapshot for the target LUN after the
reverse synchronization is started. reverse synchronization is started.

2 Copy all data a, b, c, and d to the 2 Copy incremental data d to the source LUN.
source LUN.

17 Huawei Confidential

• For more information, log in to http://support-it.huawei.com/docs/zh-cn/dorado-v6.

HyperSnap Principles - Restrictions on Feature Configuration

Feature Restriction

A source LUN of HyperSnap can be used as a source LUN of HyperClone, but a LUN of HyperSnap cannot be used as a
HyperSnap
target LUN of HyperClone.

HyperMetro A member LUN of HyperMetro can be used as a source or a target LUN of HyperClone.

HyperReplication Primary and secondary LUNs of HyperReplication can be used as a source or a target LUN of HyperClone.

SmartMigration A source or a target LUN of HyperClone cannot be used as the source or target LUN of SmartMigration.

SmartVirtualization A heterogeneous LUN cannot be used as a source or a target LUN of HyperClone.

HyperCDP A source LUN of HyperCDP can be used as a source or a target LUN of HyperClone.

18 Huawei Confidential
 Application Scenarios - Data Backup and Restoration
Source LUN Target LUN

Time point
A a ⚫ Create HyperCopy.

⚫ Synchronize data on a source LUN to a target LUN. In this

Time point case, the target LUN stores the data on the source LUN at
a a
B time point B.

Time point
C a ⚫ Data on the source LUN is lost.

Time point ⚫ Reversely synchronize data on the target LUN to the

D a a source LUN. In this case, the source LUN is restored
to the status at time point B.

19 Huawei Confidential

• HyperClone generates one or multiple copies of source data to achieve point-in-time

backup, which can be used to restore the source data in the event of data corruption.
 Application Scenarios - Data Analysis and Reproduction

Data analysis
Source Data analysis The data analysis service uses data on a target
LUN host
LUN to prevent the data analysis service and
production service from contending for
Reproducing n resources of a source LUN and affecting
pieces of data Only for data
analysis performance.

Data reproduction

Target LUN Target LUN Target LUN (n) HyperClone can create multiple copies of the same
(1) (n-1) source LUN for multiple target LUNs.

20 Huawei Confidential

• Data analysis researches on a great amount of data to extract useful information, draw
conclusions, and support decision-making. The analysis services use data on target LUNs to
prevent contention of source LUN resources between the analysis and production services,
ensuring system performance.

• HyperClone can create multiple copies of the same source LUN LUN for multiple target
LUNs.
Configuration Process
Start

Check the license.

Create a protection group.

Create a clone pair.

Create a clone consistency
group.

End

21 Huawei Confidential
 Contents

1. HyperSnap

2. HyperClone

3. HyperReplication

4. HyperMetro

22 Huawei Confidential
 Overview
⚫ As a core technology for DR and backup, HyperReplication can realize the remote data
backup and disaster recovery.

Function Purpose Benefit

This function prevents damage caused by

Remote backup and To recover service data using backup data in the remote
data loss in the case that data at the
recovery storage system after the service data is invalid.
primary site becomes unavailable.

To quickly switch service data from the primary site to This function prevents damage caused by a
Continuous service
the secondary site to protect service continuity when a service interruption upon a failure at the
support
disaster occurs. primary site.

This function prevents damage caused by

To recover data at the primary site using backup data at
DR service data loss or a long recovery duration
the secondary site after a disaster.
after a disaster.

23 Huawei Confidential

• With digitalization promoted in a wide range of industries, data has become critically
important to the efficient operation of enterprises and public institutions, and users require
increasingly higher data storage stability. Although many vendors can offer highly reliable
storage devices, irrecoverable damage caused by natural disasters to production systems
cannot be prevented.

• Note 1: A primary site is a production center that includes the primary storage system,
application servers, and links.

• Note 2: A secondary site is a backup center that includes the secondary storage system,
application servers, and links.

• Note 3: Unless otherwise specified, hosts mentioned in this document refer to application
servers.
 Introduction to DR and Backup

⚫ When the HyperReplication feature is used, two data centers work in active/standby mode. The primary site is in the service
running status, and the DR center is in the non-service running status.
⚫ For active/standby DR, when a device in data center A is faulty or even the entire data center A is faulty, services are
automatically switched to data center B.
⚫ For backup, data center B backs up only data in data center A and does not carry services when data center A is faulty.

Data center B (secondary site)

Data center A (primary site)

24 Huawei Confidential

• Recovery Point Objective (RPO): refers to the minimum time period for recovering data at
the time of a failure or disaster. Data preceding the failure or disaster at least this time
period is preserved by recovery. It also refers to the tolerable amount of lost data. RPO
indicates the timing period in asynchronous remote replication.

• Recovery Time Objective (RTO): refers to the maximum acceptable time period required to
bring one or more applications and associated data back from an outage to a correct
operational state. The RTO uses the recovery time point as the objective and ensures that
the redundancy process can take over services as quickly as possible. It also refers to the
tolerable service interruption time. RTO of asynchronous remote replication depends on
host services and fault scenarios.
HyperReplication Concepts
⚫ Question: What are HyperReplication pairs, consistency groups, synchronization, splitting,
primary/secondary switchover, data status, and writable secondary LUNs?
⚫ To implement remote backup and recovery of service data, HyperReplication involves the following
phases: creating a HyperReplication relationship, data synchronization, service switchover, and data
recovery. What do these phases mean?

25 Huawei Confidential
 Phases for Realizing Remote Backup and Recovery of Service
Data
1. Create a HyperReplication pair. 2. Synchronize data.
Primary LUN Secondary LUN Primary LUN Secondary LUN

HyperReplication HyperReplication
Link Link

WAN WAN

Pair Pair

3. Switch over services. 4. Recover data.

Primary LUN Secondary LUN Primary LUN Secondary LUN

HyperReplication HyperReplication
Link Link
WAN WAN

Pair Pair

26 Huawei Confidential

• A pair refers to the data replication relationship between a primary LUN and a secondary
LUN. In HyperReplication, data can be replicated only from the primary LUN to the
secondary LUN through a remote replication link. Before the data synchronization, a pair is
necessary for data replication between a primary LUN and a secondary LUN. The pair
relationship establishment is similar to the process of dialing a number to set up a
connection between two phones. In configuring HyperReplication, a pair consists of a
primary LUN on the primary storage system and a secondary LUN on the secondary storage
system.

• A pair for HyperReplication: A pair is formed after HyperReplication is created. Multiple

associated pairs can form a consistency group.

• Data synchronization: After the HyperReplication pair is created, data on the primary LUN is
initially synchronized to the secondary LUN in full mode by manual or automatic manner.
The incremental synchronization of data on the primary LUN is periodically performed on
the secondary LUN.

• Service switchover: Users can determine whether a primary/secondary switchover can be

performed by viewing the data status. A new pair is formed by the primary-secondary
switchover.

• Data recovery: Data on the secondary storage system is synchronized to the primary
storage system. The old pair is recovered by the primary-secondary switchover.
 Running Status of a Pair
⚫ By viewing the running status of a pair, you can perform synchronization, splitting, and primary/secondary switchover
operations on HyperReplication in time. After performing an operation, you can view the running status of the pair to check
whether the operation is successful.

Running Status Description

Normal Indicates that data synchronization between the primary and secondary LUNs is complete.
Indicates that data replication between the primary and secondary LUNs is suspended. For the need of services, a pair of primary and secondary
Splitting
LUNs is manually split. As a result, the running status of the pair changes to the splitting status.
Indicates that if the pair relationship between the primary and secondary LUNs is interrupted because the link used by HyperReplication is down
Interrupted
or the primary or secondary LUN of HyperReplication is faulty. The pair running status is interrupted.
Indicates that if HyperReplication requires to be restored using a manual policy after the fault caused by a pair interruption is rectified, the pair
To be recovered running status changes to the to-be-recovered status. This status indicates that users need to manually synchronize data between the original
primary LUN and the secondary LUN to restore the pair.
Indicates that if the original attributes of a primary or secondary LUN change when a pair is interrupted (for example, when the HyperReplication
Invalid link is down, the pair is deleted on the primary or secondary end), the pair running status changes to the invalid status because the primary and
secondary system configurations become inconsistent.
When the primary LUN is synchronizing data to the secondary LUN, the secondary LUN cannot be read or written. If a disaster occurs, data on the
Synchronizing secondary LUN cannot be used for service recovery. When the secondary LUN is in the complete status, data on the secondary LUN can be used
for service recovery.

27 Huawei Confidential

• Description: If Initial Synchronization is set to The data on primary and secondary resources
is consistent and data synchronization is not required, a newly created remote replication
pair is in the Normal state.
 Principles of Asynchronous Remote Replication
⚫ Based on the known asynchronous
Host
remote replication, try to draw a
1 2 DCL
schematic diagram of synchronous 3 6
remote replication. RM
1 2

Primary 5
Secondary
LUN Cache

Primary 5 Secondary
LUN LUN
4 4
Snapshot of the HyperReplication Link Snapshot of the
primary LUN secondary LUN
7 WAN 7

Snapshot of the primary Snapshot of the

LUN secondary LUN

DCL
Secondary Storage System
Primary Storage System

28 Huawei Confidential

• Currently, Dorado V6 6.0.0 does not support synchronous remote replication.

• Key points: the read and write process of asynchronous remote replication

• A host delivers an I/O request to a cache.

• DCL records differences and writes data to the primary LUN.

• Snapshots of primary and secondary LUNs are activated.

• Incremental data of the primary LUN is synchronized to the secondary LUN.

• The storage system checks whether data differences are eliminated by DCL based on the
write operation.

• The snapshot is stopped.

• Logs of DCL are stored in four disks randomly selected by the storage system. Operation
and execution logs are stored in the database of the system.
 HyperReplication Service Switchover
Production Host Production host Standby host
Standby Host
√

Services run by the Read-only

Read-only
production host
Primary Disconnected Secondary
Primary Replication Secondary
LUN Link LUN
LUN Link LUN
WAN WAN
Secondary storage Primary storage Secondary storage
Primary storage
1. Normal production at the primary site 2. A disaster at the primary site

Production host Standby host

√

Services
Read/Write taken
over by
Primary Disconnected Secondary the 3. Service taken over by the secondary site
LUN Link LUN standby
host
WAN
Primary storage Secondary storage

29 Huawei Confidential

• When the primary site of HyperReplication suffers a disaster, the secondary site can quickly
take over services to protect service continuity.

• HyperReplication not only implements remote data backup but also recovers services as
soon as possible in the case of a disaster to maintain service continuity. You must consider
the following two indicators when switching over services.

• Requirements for running services on the secondary storage system:

▫ Before a disaster occurs, data in the primary LUN is consistent with that in the
secondary LUN. If data in the secondary LUN is incomplete, services may fail to be
switched.

▫ Services on the production host have also been configured on the standby host.

▫ The secondary storage system allows a host to access a LUN in a LUN group mapped
to the host.

• When a disaster occurs, the primary site is invalid, and the HyperReplication links between
the primary and secondary LUNs are down. In this case, the administrator needs to
manually set read/write permission of the secondary LUN to writable mode to implement
the service switchover.
 HyperReplication Data Recovery
Production host Standby host Production host Standby host
√
Data
recovered
Services run
by the
Secondary Primary standby
Link Secondary New Primary
LUN LUN host.
Recovery LUN Link LUN

WAN WAN
Primary storage Secondary storage Primary storage Secondary storage

1. Disaster recovery at the primary site. 2. Data recovery at the primary site

Production host Standby host

√
Services
run by the 3. Service recovery at the primary site
productio Unreadable
n host Primary Unwritable
Replication Secondary
LUN
Link LUN
WAN
Primary storag Secondary storage

30 Huawei Confidential

• After the primary site of HyperReplication fails, the secondary site temporarily takes over
services of the primary site. When the primary site recovers, services are switched back.

• After the primary site recovers from a disaster, it is required to rebuild a HyperReplication
relationship between the primary and secondary storage systems and use data on the
secondary site to recover data on the primary site.

• In an asynchronous remote replication scenario, the storage system performs data

synchronization multiple times until the data difference between the primary and
secondary LUN is small. Then, the storage system stops services and performs the last
synchronization. In this way, no data is lost and service downtime is minimized.
 Functions of a Consistency Group
Primary LUN 01 CG 01 Secondary LUN 01 Primary LUN 01 CG 01
Secondary LUN 01
HyperReplication
HyperReplication Task 01
Task 01
Primary LUN 02 Secondary LUN 02 Primary LUN 02 Secondary LUN 02
HyperReplication
HyperReplication Task 02
Task 02
Primary LUN 03 Secondary LUN 03 Primary LUN 03 Secondary LUN 03
HyperReplication
HyperReplication Task 03
Task 03

1. Create a consistency group. 2. Stop replication tasks for the consistency group.

Primary LUN 01 CG 01 Secondary LUN 01

HyperReplication
Task 01
Primary LUN 02 Secondary LUN 02

HyperReplication 3. Ensure data validity for the consistency group.

Task 02
Primary LUN 03 Secondary LUN 03
HyperReplication
Task 03

32 Huawei Confidential

• In medium- and large-size database applications, data, logs, and change records are stored
in different but associated LUNs in the storage system. Data association among these LUNs
is ensured by upper-layer services of hosts at the primary site. This association must be
maintained when these LUNs are replicated to the secondary site. If the association is
unavailable, DR and backup data of the secondary storage system cannot be used to
recover services. In this case, HyperReplication pairs of these LUNs can be added to a
consistency group. This slide introduces the consistency group function of HyperReplication
by comparing and analyzing data invalidity of a storage system with a consistency group
and a storage system without a consistency group.

• By creating a consistency group, users can perform synchronization, splitting, and primary-
secondary switchovers for a single HyperReplciation pair or for multiple HyperReplication
pairs in a consistency group. Note the following when creating a HypeReplication
consistency group:

▫ HyperReplication pairs can only be added to a consistency group on the primary

storage system, and all secondary LUNs of HyperReplication pairs must lie in the
same remote storage system.

▫ LUNs of different HyperReplication pairs in a consistency group can lie in different

working controllers.

▫ HyperReplication pairs in a consistency group must belong to the same replication

mode.
 Typical Application Scenarios
Analysis Item Central DR and Backup Geo-redundancy

Backup data is managed centrally so that data analysis

Three data centers are deployed in two cities to perform real-
and data mining can be performed without affecting
time backup and remote backup concurrently.
services.
Service data is backed up to an intra-city DR center in real time
When a disaster occurs at any service site, the central
through a high-speed link.
DR and backup site can quickly take over its services
Scenario feature After data in the primary site is invalid, services are quickly
and recover data, achieving unified service data
switched to the intra-city DR center.
management.
If a disaster damages the primary site and the DR center in the
HyperReplication mode can be selected for a service
same city, an inter-city DR center takes over services and
site flexibly based on the distance between the service
implements DR.
site and the central DR and backup site.

HyperReplication Intra-city: asynchronous remote replication

Asynchronous remote replication
mode Inter-city: asynchronous remote replication

Maximum distance
Asynchronous remote replication: no restriction Asynchronous remote replication: no restriction
for DR and backup

33 Huawei Confidential

• This solution belongs to Huawei Business Continuity and Disaster Recovery Solution. In the
Disaster Recovery Data Center Solution (Geo-Redundant Mode), three data centers coexist.
The continuity of core services can be ensured when two data centers are damaged,
remarkably improving availability of the disaster recovery solution. The primary site, intra-
city DR center, and inter-city DR center are the three data centers in the solution.

▫ Primary site: provides services externally.

▫ Intra-city DR center: locates at a place dozens of kilometers away from the primary
site. Fiber Channel network–based direct connection is recommended. Synchronous
replication is implemented, which is level-1 disaster recovery protection of the geo-
redundant solution.

▫ Inter-city DR center: locates at a place hundreds of or thousands of kilometers away

from the primary site to cope with regional disasters. Periodic asynchronous
replication DR is implemented, which is level-2 DR protection of the geo-redundant
solution.

• Geo-redundancy can be used when you want to take multi-protection for data and services
in the primary site.

▫ Service DR: If a disaster occurs in the primary site, the intra-city DR center can
quickly take over services. If disasters occur in both the primary site and the intra-city
DR center, the inter-city DR center can restore production services using data copies.
In this way, service continuity is ensured to the maximum.
 Central DR and Backup Scenario
⚫ Central DR and backup refer to backing up service data from different places to the same site for centralized management.
Service data at multiple service sites is centrally backed up to and managed at the central DR and backup site. When a
disaster occurs, the central DR and backup site can take over services from the service site and recover data.

Secondary LUN
Primary LUN 02 HyperReplication 02 02

WAN

Host
Service site 02 Asynchronous

......
......

......
Secondary LUN
Primary LUN n HyperReplication n n

WAN

Service site n Asynchronous Central DR and

backup site

34 Huawei Confidential

• Functions of HyperReplication are as follows:

• Collects data from n (a number) service sites to the central DR and backup site, takes a
snapshot of a secondary LUN, and maps the secondary LUN to a host. The snapshot can be
used for data analysis and data mining.

• Allows the central DR and backup site to take over services temporarily when any service
site (01 to n) has a disaster and switches back services when the site recovers.

• Performs replication tasks (using the asynchronous remote replication mode) from service
site 02 to the central DR and backup site, which is 500 km away from service site 02.
 Realizing DR Used with BCManager eReplication
⚫ BCManager eReplication is a DR management software specially designed for Huawei typical DR solutions. It provides a
visualized and process-based platform for simple and fast operations and monitoring based on data consistency, HyperSnap,
and HyperReplication technologies.
Primary site DR center

BCManager BCManager
Production host eReplication
eReplication
Server Agent
BCManager
eReplication
Agent

Standby host

BCManager
eReplication
Host

HyperReplication

Storage system Storage system

35 Huawei Confidential

• Deploy the BCManager eReplication host in the DR center and install BCManager
eReplication Server on the host. Install BCManager eReplication Agent on the service and
standby host in the primary site.
Configuration Process
Start
Required
Check the license.
Optional
Create a logical port.

Manage a route.

Create an authentication user.

Add a remote device.

Create a protection group.

Create a HyperReplication
pair. Create a remote replication
consistency group.

End

36 Huawei Confidential
 Contents

1. HyperSnap

2. HyperClone

3. HyperReplication

4. HyperMetro

37 Huawei Confidential
Overview
⚫ HyperMetro is also called active-active feature. Two data centers are backups for each
other in the running status.
⚫ If a device is faulty in a data center or even the entire center is faulty, the other data
center will automatically take over services, solving the problems of traditional DR centers
in switchover. This ensures high data reliability and service continuity, and improves the
resource utilization of the storage system.

38 Huawei Confidential
 Working Principles of HyperMetro
Data center A Data center B

How to perform
arbitration when a
Oracle RAC cluster/VMware fault occurs?
vSphere cluster/
FusionSphere cluster
......

WAN
FC/IP SAN SAN SAN
FC/IP

Production
Real-time data synchronization
Production
storage storage

IP network IP network

Quorum device

39 Huawei Confidential

• Active-active storage design of HyperMetro:

▫ A-A structure: Active-active LUNs are readable and writable in both data centers and
data is synchronized in real time.

▫ High-reliability design: The double-arbitration mechanism and cross-DC bad block

repair improve system reliability.

▫ High-performance design: Multiple performance tuning measures are provided,

reducing latency of interactions between two data centers and improving service
performance by 30%.

▫ Flexible scalability design: Heterogeneous storage arrays, snapshot, and remote

replication interworking are supported. The Disaster Recovery Data Center Solution
(Active-Active Data Center) can be expanded to the Disaster Recovery Data Center
Solution (Geo-Redundant).
Quorum Mode
⚫ If the link between two data centers is down or one data center is faulty, data cannot be
synchronized between the two data centers in real time. In this case, only a HyperMetro
pair or a site of HyperMetro consistency group can continue providing services. For data
consistency, HyperMetro adopts an arbitration mechanism to determine service priority in
data centers.
⚫ HyperMetro provides two quorum modes:
 Static priority mode: applied to scenarios where no quorum server is configured.
 Quorum server mode (recommended): applied to scenarios where a quorum server is
configured.

40 Huawei Confidential
 Static Priority Mode
Fault Type Result
A HyperMetro pair is in the to-be-synchronized status.
A link between storage
LUNs in data center A continue providing services while LUNs in data center
systems is down.
B stop providing services.
A HyperMetro pair is in the to-be-synchronized status.
Host Data center B is faulty. LUNs in data center A continue providing services while LUNs in data center
B stop providing services.
A HyperMetro pair is in the to-be-synchronized status.
Data center A is faulty. LUNs in data center A cannot be accessed and LUNs in data center B stop
providing services.
A HyperMetro replication link
A HyperMetro pair is in the to-be-synchronized status.
HyperMetro is down and a link between a
LUNs in data center A continue providing services while LUNs in data center
host and data center B is
B stop providing services.
down.
Data center B is faulty and A HyperMetro pair is in the to-be-synchronized status.
the link between the host LUNs in data center A continue providing services while LUNs in data center
Data center A Data center B and data center B is down. B stop providing services.
Links between the host and
A HyperMetro pair is in the normal status.
data centers A and B are
A host fails to access LUNs in both data center A and B.
concurrently down.

41 Huawei Confidential

• When no quorum server is configured or the quorum server is inaccessible, HyperMetro

works in static priority mode. When an arbitration occurs, the preferred site wins the
arbitration and provides services.

▫ If the links between the storage arrays are down or the non-preferred site breaks
down, LUNs at the preferred site continue providing Metro services and the LUNs at
the non-preferred site stop.

▫ When the preferred site is faulty, the non-preferred site cannot automatically take
over HyperMetro services. As a result, the services stop and you need to forcibly start
the non-preferred site to provide services for the host.
 Quorum Server Mode

Host Fault Type Result

A HyperMetro pair is in the normal status.
The quorum server is faulty. LUNs in data center A and data center B continue providing
services.
A link between a storage system and A HyperMetro pair is in the normal status.
the quorum (example of the storage LUNs in data center A and data center B continue providing
system in data center A) is down. services.
HyperMetro
A HyperMetro pair is in the to-be-synchronized status.
A storage system is faulty (example of
LUNs in data center A are invalid, but LUNs in data center B
the storage system in data center A).
continue providing services.

Data center A Data center B A HyperMetro pair is in the to-be-synchronized status.

A link between storage systems is
LUNs in data center A continue providing services while LUNs in
down.
data center B stop providing services.

A storage system and the quorum

server (example of the storage system A HyperMetro pair is in the to-be-synchronized status.
in data center A) are concurrently Data center A is faulty and LUNs in data center B stop services.
Quorum server faulty.

42 Huawei Confidential

• An independent physical server or VM is used as the quorum server. It is recommended

that the quorum server be deployed at a dedicated site that is in a different fault domain
from the two DCs. In this way, when a disaster occurs in the single data center, the quorum
server still works.

• In quorum server mode, in the event of a DC failure or disconnection between the storage
systems, each storage system sends an arbitration request to the quorum server, and only
the winner continues providing services. The preferred site takes precedence in arbitration.
 Dual-Write Principle

Host

Write I/O
1 5
Cross-site active-active cluster

Array A Array B
2 HyperMetro LUN

4
4 3 3

HyperMetro LUN HyperMetro LUN

44 Huawei Confidential

• Dual-Write Principle

▫ A host delivers an I/O write request.

▫ A distributed lock is applied.

▫ The local storage array writes data to the local cache and sends the write request to
the remote storage array.

▫ The data is written into the local cache and remote storage array successfully. The
remote storage array returns a write success message to the local storage array.

▫ The data is written into both the local and remote storage arrays successfully and the
storage arrays return a write success message to the host.
 Strong Data Consistency

Data center A Data center B

Host Application cluster Host

Cross-site active-active cluster

Mutual exclusion of HyperMetro LUNs' distributed locks

Array A Array B

HyperMetro LUN HyperMetro LUN

45 Huawei Confidential

• Data consistency at the application layer: Cross–data center databases, applications

deployed in a cluster, and shared storage architecture

• Dual-write of I/Os, real-time data consistency

• Data consistency at the storage layer

• Dual-write of I/Os ensures consistent data.

• In normal conditions, the application I/Os that are delivered are concurrently written into
both storage arrays, ensuring data consistency between the two storage arrays.

• Differential data recording upon the breakdown of a single storage array

• If a storage array breaks down, data is written into the other storage array that is working
properly and data changes are recorded in a data change log (DCL). After the storage array
is recovered and connected to the system again, the data changes in the DCL are written
into the storage array in incremental mode.

• Distributed lock management (DLM): Only one host is allowed to write data to a storage
address at a time when multiple hosts are accessing the address simultaneously. This
ensures data consistency.
Solution Extensibility Design
Disaster recovery solution HyperMetro data center solution
Site A Site B Site A
Upgraded Site B
architecture and
uninterrupted
services
Synchronous/
asynchronous HyperMetro
replication

Geo-redundant solution
Data Data center Data center C
center A B

46 Huawei Confidential
Typical Application Scenarios
Industry Feature

With the development of hospital services, the growing numbers of beds and new outpatient buildings pose higher
requirements on service continuity.
Once critical departments such as out-patient, in-patient, and electronic medical record (EMR) are interrupted, medical
Healthcare
treatment will be delayed and hospitals will suffer from great economic loss and inestimable damage to their reputation. In
addition, an out-patient building is close to an in-patient network information center in the same hospital, and two hospitals in
the same city are physically close to each other. HyperMetro can meet their requirements.

In the finance industry, banking services, 24-hour ATM services, POS services, and e-bank services are developing quickly as
bank services develop. These services require that banking systems process around-the-clock services.
Finance For reliability and stability, banks require a solution to store for reused and meet their service construction requirements (RPO =
0, RTO = 0) to ensure business continuity. Service interruptions damage banks' reputation and pose huge pressure on technical
departments.

In the social security industry, service continuity requirements are high. Monthly settlement and year-end carry-over require
24/7 online operation. Otherwise, people's livelihood problems may occur. For example, pension cannot be paid in time, and
medical insurance cannot be settled in time.
Social security
HyperMetro is applicable to social insurance application scenarios including the basic information management, social
insurance card service, labor relationship, public services, public resource management, employment, and social insurance
management.

47 Huawei Confidential
Configuration Process
Start
Required
Prepare the configuration. Optional

Configure the switches.

Configure quorum server

software. Check the license.

Add a remote device.

Create a quorum server.

Configure HyperMetro.
Create a HyperMetro domain.

Create a HyperMetro pair.

Configure a multipathing
policy for hosts. Create a HyperMetro
consistency group.

End

48 Huawei Confidential
 Summary

HyperSnap

HyperClone
Storage Data Protection Definitions, principles,
and configuration
Technologies and Applications
processes
HyperReplication

HyperMetro

49 Huawei Confidential
 Quiz

1. (True or false) A source LUN can form multiple HyperClone pairs with different target LUNs. A target
LUN can be added to only one HyperClone pair.

2. Which of the following statements are correct? ( )

A. Synchronous replication synchronizes data in real time to maximize data consistency and minimize data loss in
the event of a disaster.

B. Asynchronous replication performs periodic data synchronization, minimizing service performance

deterioration caused by data transmission latency.

C. HyperReplication does not require a license.

D. Synchronous replication uses snapshots for replication.

50 Huawei Confidential

• Answers:

▫ T

▫ AB
 Quiz

1. Which of the following statements about consistency groups are correct? ( )

A. If multiple LUNs at the primary end have write dependency relationships, replication consistency
groups must be configured to ensure that the secondary LUNs also have write dependency
relationships.

B. A consistency group is only used to manage the replication relationship between multiple LUNs.

C. If one member in a consistency group is faulty, all members in the consistency group are
unavailable.

51 Huawei Confidential

• Answers:

▫ AC
 More Information

Enterprise Huawei
technical support enterprise
app business app

52 Huawei Confidential

• Huawei training app

▫ Contains a large number of Huawei certified high-quality learning videos.

• Enterprise technical support app

▫ Covers all popular product documents, cases, and bulletins of Huawei. Users can
quickly query commands, alarms, and spare parts, and scan the QR to view the
device information and simple as well as intuitive video guide. It provides
uninterrupted enterprise technical support.

• Huawei enterprise business app

▫ Provides one-stop mobile ICT portals for customers and partners to understand
Huawei's comprehensive product and solution information in the enterprise ICT field
anytime and anywhere.
 Recommendations

⚫ Huawei official websites

 Enterprise business: https://e.huawei.com/en/
 Technical support: https://support.huawei.com/enterprise/en/index.html
 Online learning: https://www.huawei.com/en/learning

⚫ Popular tools
 HedEx Lite
 Network Document Tool Center
 Information Query Assistant

53 Huawei Confidential

• Popular tools:

▫ HedEx Lite: Huawei product document management tool, which allows users to
browse, search for, update, and manage product documentation.

▫ eStor: A graphic storage simulation platform. Through simulation of Huawei

OceanStor all-flash storage devices, the platform helps ICT practitioners and
customers quickly get familiar with Huawei storage products, and understand and
master their operations and configurations.

▫ Network Documentation Tool Center: The documentation tool for network products
is a good assistant for bidding support, network planning, project delivery, and
upgrade and maintenance.

▫ Information Query Assistant: It provides commands and alarm information queries

for Huawei products.
Thank you. 把数字世界带入每个人、每个家庭、
每个组织，构建万物互联的智能世界。
Bring digital to every person, home, and
organization for a fully connected,
intelligent world.

Copyright©2020 Huawei Technologies Co., Ltd.

All Rights Reserved.

The information in this document may contain predictive

statements including, without limitation, statements regarding
the future financial and operating results, future product
portfolio, new technology, etc. There are a number of factors that
could cause actual results and developments to differ materially
from those expressed or implied in the predictive statements.
Therefore, such information is provided for reference purpose
only and constitutes neither an offer nor an acceptance. Huawei
may change the information at any time without notice.