1

Universal Replicator overview

With Hitachi Universal Replicator (UR), you create and maintain a remote copy of a data volume on a Hitachi Virtual Storage Platform (VSP) storage system.

This guide provides instructions for planning, configuring, operating, maintaining, and troubleshooting a Universal Replicator system.

□

Universal Replicator software

□

How Universal Replicator works

□

Hardware and software components

□

Reference Information

Universal Replicator overview

Hitachi Virtual Storage Platform Hitachi Universal Replicator User Guide

1–1

Universal Replicator software

With Universal Replicator, you copy application data to a secondary VSP system at a remote location. The remote volume is an asynchronous blockfor-block copy of the local storage volume. The copied data is consistent with local data and therefore available for recovering the local volume if the need arises.

Universal Replicator is designed to support a secondary site hundreds and even thousands of miles from the local site, making recovery from regionwide disasters possible.

Universal Replicator is also designed to limit impact on the local system.

Updates sent from a host to the primary production volume on the local system are copied to a local journal volume. The remote system “pulls” data from the journal volume across the communication link to the backupvolume, called the secondary volume. The local system is free to perform its role as a transaction processing resource rather than as a replication engine.

Note: The ”local” system is also referred to as the “primary” system.

The “remote” system is also referred to as the “secondary” system.

The production volume on the primary system receives and stores the data sent from the host. This volume is commonly referred to as the “primary volume”, or “P-VOL”.

The backup volume on the secondary system stores the data copied from the primary system, and is commonly referred to as the “secondary volume”, or “S-VOL”.

How Universal Replicator works

Remote replication occurs using journal volumes on the local and remote systems.

1–2

Figure 1-1 Basic sequence in Universal Replicator operations

• The journal volume on the local system is called the “master journal volume”.

Universal Replicator overview

Hitachi Virtual Storage Platform Hitachi Universal Replicator User Guide

• The journal volume on the remote system is called the “restore journal volume”.

Replication occurs in the following sequence:

1. Journal obtain - When the host sends an update to the primary volume

(P-VOL), the primary (local) system’s journal-obtain function triggers a copy of the updated data to the master journal volume.

The host assigns write-sequence numbers to the data sent to the master journal volume.

Write-sequence numbers and other metadata attached to journal data ensure consistency with the data in the P-VOL.

2. Journal copy - Data is copied from the master journal to the restore journal.

When the master journal has data, the data is transferred to the restore journal. When data transfer is complete, master journal data is discarded.

Data copy to the restore journal is initiated by the read-journal command issued by the remote system.

Data copy occurs on a continual basis unless there is no data in the master journal. The request for data from the remote system is repeated as soon as the previous read operation is completed.

Journal data is removed from the master journal only when the primary system receives sequence numbers for the data from the restore journal.

3. Journal-restore - The secondary volume (S-VOL) is updated with changed data from the restore journal.

Data is copied to the S-VOL according to the write sequence numbers, ensuring data consistency.

When journal-restore is completed, the data in the restore journal is discarded.

Performance is affected because of journal-to-journal copying. Usage rates

are also affected. See Read and write I/O during remote copy on page 1-8

for more high level information on Universal Replicator operations.

Hardware and software components

A typical configuration consists of a VSP system or externally attached storage system on both local and remote sites, a host or hosts connected to the systems, Universal Replicator software on both systems, data path connections, and interface tools for configuring and managing Universal

Replicator.

• The local and remote VSP systems are connected using dedicated fibrechannel data paths that can include fibre-channel switches. Data paths are routed from the fibre-channel ports on the primary system to the ports on the secondary system.

• The host is connected to the VSP using a fibre-channel or fibre-channelover-Ethernet (FCoE) target port.

Universal Replicator overview

Hitachi Virtual Storage Platform Hitachi Universal Replicator User Guide

1–3

• Storage Navigator, whose GUI is used to view and operate Universal

Replicator, is connected via a management LAN.

A Universal Replicator system consists of the following:

• P-VOLs and S-VOLs on the local and remote VSP

• Master and restore journal volumes on the local and remote VSP

• Master and restore journals on the local and remote VSP

The master journal consists of the primary volumes and master journal volumes.

The restore journal consists of the secondary volumes and restore journal volumes.

Management software consists of:

• Storage Navigator graphical user interface (GUI)

• Command Control Interface (CCI)

Universal Replicator components are illustrated in the following figure and described in greater detail in the following topics.

Figure 1-2 Universal Replicator components

VSP storage systems

Universal Replicator is operated using two VSP systems, one at the primary site and one at the secondary site. The primary system consists of the main control unit (MCU) and the service processor (SVP). The secondary system consists of the remote control unit (RCU) and the SVP.

1–4

Universal Replicator overview

Hitachi Virtual Storage Platform Hitachi Universal Replicator User Guide

• The primary system communicates with the secondary system over dedicated fibre-channel remote copy connections.

• Each VSP system can function simultaneously as a primary and secondary system.

Main and remote control units

The primary and secondary systems are often referred to as the MCU

(primary system) and RCU (secondary system). MCU is the main control unit, RCU is the remote control unit.

The MCU control the primary storage volume (P-VOL) and the following operations:

• Host I/O write to the P-VOL

• P-VOL data copy to the master journal

• Initial copy and update copy between the P-VOL and secondary volume

(S-VOL).

The RCU control the secondary storage volume (S-VOL) and the following operations:

• Journal commands to the MCU.

• Journal data copy from the master to the restore journal

• Restore journal data copy to the S-VOL

• Pair status management and configuration (for example, rejecting write

I/Os to the S-VOLs).

Pair volumes

Original data is stored in the P-VOL and the remote copy is stored in the S-

VOL. The pair can be paired, split, re-synchronized, and returned to the unpaired (called “simplex”) state. When synchronized, the volumes are paired; when split, new or changed data sent to the P-VOL is not copied to the S-VOL. When re-synchronized, changed data is copied to the S-VOL. If a disaster occurs, production operations can be transferred to the S-VOL.

When the primary site is functional again, operations can be transferred and data can be copied back to the P-VOL.

The P-VOL remains available to the host for read and write I/O operations.

The secondary system rejects write I/Os for the S-VOL, unless the writeenable option is specified. Then, write I/O is allowed to the S-VOL while the pair is split. In this instance, S-VOL and P-VOL track maps keep track of differential data and use it to re-synchronize the pair.

Journal volumes

For Universal Replicator operations, journal volumes are required on the primary and secondary systems.

• Updates to the P-VOL are copied to the master journal volume in the

primary system. See the illustration in Journals on page 1-6

.

Universal Replicator overview

Hitachi Virtual Storage Platform Hitachi Universal Replicator User Guide

1–5

• Master journal data is copied to the restore journal volume on the secondary system.

• Journal volumes can have different volume sizes and different RAID configurations.

• Journal data is stored sequentially and separately in each journal volume in the same journal.

For information on planning journal volumes, see Sizing journal volumes on page 3-4

.

Journals

Journals help you manage data consistency between multiple P-VOLs and

S-VOLs. A journal consists of two or more data volumes and journal volumes.

You use journals to create multiple pairs and to split, resynchronize, and release multiple pairs. Journals are required on the primary and secondary systems.

Each data volume and its associated journal volume reside in the same journal.

• The master journal contains master journal volumes and is associated with the P-VOL.

• The restore journal contains restore journal volumes and is associated with the S-VOL

Each pair relationship between journals is called a "mirror". A mirror ID identifies a pair relationship between journals. When the pair is created, it is assigned a mirror ID.

Figure 1-3 Journals

1–6

Universal Replicator overview

Hitachi Virtual Storage Platform Hitachi Universal Replicator User Guide

Data path

The physical transmission link between the local and remote systems is called the data path. Universal Replicator commands and data are transmitted through the fibre-channel data path and switches. The data path is connected to the primary and secondary systems through two types of fibre-channel ports, Initiator and RCU Target ports. Universal Replicator requires paths in both directions. More specifically, it requires paths with

Initiator ports in the MCU connected to RCU Target ports in the RCU, and paths with Initiator Ports in the RCU connected to RCU Target ports in the

MCU.

One data path connection is required. It is recommended that you use two or more independent connections to provide hardware redundancy. A maximum of eight paths can be used.

For more information, see

Planning the data path on page 4-1

.

Consistency groups and journals

Journals are used in Universal Replicator to guarantee data consistency across multiple pairs. Consistency groups are used in other replication software for the same purpose. As a best practice, you can assign CCI consistency group numbers as journal numbers. See “journals” in

System requirements on page 2-2

and

Multiple journals per CCI consistency group on page 3-14

for more information.

Storage Navigator

Storage Navigator provides a GUI and command line interface for accessing and managing the storage system, including Universal Replicator.

Storage Navigator communicates with the SVP of each system over defined

TCP/IP connections.

Command Control Interface (CCI)

CCI provides a command line interface for accessing and managing the storage system, including Universal Replicator. You can perform the same

Universal Replicator operations with CCI as you can with Storage Navigator.

In addition, you can automate pair operations using scripts.

If you want to use CCI but it is not installed, contact your Hitachi Data

Systems account team.

Reference Information

The following topics describe UR copy operations and other features.

Overview of copy operations

initial and update copy operations including the underlying operations, such as journal processing and differential data management.

Universal Replicator overview

Hitachi Virtual Storage Platform Hitachi Universal Replicator User Guide

1–7

Initial copy operation

The initial copy is executed when the primary system copies all the data in sequence from the P-VOL directly to the S-VOL. Though journal volumes are not used during the initial copy, the copy data in this operation is referred to as “base journal data”.

• Creating or resynchronizing two or more pairs within the same journal results in the base journal data being copied to the respective S-VOLs, one at a time. This extends the time required for all the operations to be completed.

• An initial copy operation can be performed to establish the pair — with no data copied between the volumes. This can be done when data in the

P-VOLs and S-VOLs are identical.

• Universal Replicator pair data can also be copied using a TrueCopy initial copy operation. Doing this reduces the time to complete the copy operation. See

Planning pair volumes on page 3-8

for more information.

Update copy operation

When a host has new or changed information, the following occurs in the primary system:

• The update is written to the P-VOL

• The update is copied to the master journal along with metadata that includes sequence and other consistency information.

The remote system issues the read-journal command (independent of host I/O activity). At this time, the following occurs:

Any data in the master journal is sent to the restore journal.

The updated data is copied to the S-VOL.

Journal data on the primary and secondary systems is discarded when data consistency is established in the copy.

Note: Journal data is transferred using special I/O operations initiated by the secondary system, called RIO (remote I/O). RIO provides the most efficient type of data transfer. Make sure that your channel extenders are capable of supporting RIO. Contact Hitachi Data Systems

Support Center for more information.

If an update copy operation fails, the remote system suspends the affected pair or all pairs in the journal, depending on the type of failure.

The suspended pair or journal returns to Paired status when the primary and secondary systems are re-synchronized.

Read and write I/O during remote copy

The primary system reads from the P-VOL when it receives a read I/O command. If the read fails, the redundancy provided by RAID-1 or RAID-5 technology recovers the failure. The primary system does not read the S-

VOL for recovery.

1–8

Universal Replicator overview

Hitachi Virtual Storage Platform Hitachi Universal Replicator User Guide

When a primary system receives a write I/O command for a P-VOL in PAIR status, the system performs the write operation and performs the update copy operation. The write operation completes independently of the update copy operations on the S-VOL.

The secondary system updates the S-VOL according to the write sequence number in the journal data. This maintains data consistency between P-VOL and S-VOL.

If the P-VOL write operation fails, the primary system reports a unit check and does not create the journal data for this operation. As mentioned, if the update copy operation fails, the secondary system suspends either the affected pair or all Universal Replicator pairs in the journal, depending on the type of failure. When the suspended pair or journal is resumed, the primary and secondary systems negotiate the resynchronization of the pairs.

During normal operations, the secondary system does not allow S-VOLs to be online (mounted). Therefore, hosts cannot read from and write to S-

VOLs. However, if the S-VOL write option is enabled, write access to an S-

VOL is allowed while the pair is split. The pair must be split from the primary system for the option to take effect.

To reduce the overhead associated with remote copy activities and to maximize rate of data transfer, the VSP uses a special write command for initial and update copy operations. This command transfers the control parameters and the fixed-block architecture (FBA) format data for consecutive updated records in a track using a single write operation. It eliminates the overhead required for performing FBA-to-count-key-data

(CKD) and CKD-to-FBA conversions.

Differential data management

Differential data is the data that is changed in the P-VOL when a pair is split or suspended and that is not reflected in the S-VOL. This data is stored in a track bitmap. When the pair is resynchronized, the primary system merges the P-VOL and S-VOL bitmaps, and the differential data is copied to the S-

VOL.

The number of bitmap areas affects the maximum possible number of pairs that can be created in the system.

S-VOL write option

When splitting a pair, you can set an option allowing write I/O to the S-VOL.

When you resynchronize a split pair whose S-VOL is write-enabled, the secondary system sends the S-VOL track bitmap to the primary system, which merges the P-VOL and S-VOL bitmaps to determine which tracks are out of sync. This ensures proper resynchronization of the pair.

Universal Replicator overview

Hitachi Virtual Storage Platform Hitachi Universal Replicator User Guide

1–9

Pair status

Every pair operation results in a change in pair status. In addition, when you want to perform an operation, the pair must have a specific status in order to for the operation to run. You will monitor pair status to ensure that you can perform the desired operation, and to ensure that an operation completed successfully.

The following provides a brief description of the pair statuses. For complete

details, see Pair status definitions on page 7-2 .

• SMPL: A volume that is not assigned to a pair is in simplex status, SMPL.

• COPY: When copy processing is started, the primary system changes the status of the P-VOL and S-VOL to COPY.

• PAIR: When the initial copy processing is complete, the primary system changes the status of both data volumes to PAIR.

• PSUE: When a pair is suspended due to an error condition, the primary system changes the P-VOL and S-VOL status to PSUE (if the path status is normal).

• PSUS:

When a pair is split by the user (pairsplit-r), the primary or secondary system changes the status of the P-VOL and S-VOL to PSUS (if the path status is normal).

If a pair is split from the secondary system, it changes the S-VOL status to PSUS. The primary system detects the split (if path status is normal) and changes the P-VOL status to PSUS.

1–10

Universal Replicator overview

Hitachi Virtual Storage Platform Hitachi Universal Replicator User Guide