Detailed description of HP Remote snapshot copy volume

About the remote snapshot copy Function

Remote snapshot replication is a disaster recovery authorization function. This function asynchronously copies block-level data from volumes on the Local Storage System in batches) to volumes on the same system or different independent systems. Different systems can be located at the same site or different sites as the first system.

The typical replication configuration involves the following physical and logical components:

• Host connected to the Local Storage SystemTo connect to the remote storage system through the FC or iSCSI port network.

• Remote System definition. The management object on the local system enables the local system to communicate with the MC in the remote system and exchange data.

• Copy Group. Related Major volumes, which can be copied and normally hosted on two physically separated or geographically separated storage systems. These volumes are also called copy volumes.

• Master volume. This volume is a data source in the replica set and can be mapped to the host. For disaster recovery purposes, if the primary volume is offline, the Standby volume is specified as the primary volume. The primary volume exists on the primary virtual disk in the primary system.

• Backup volume. This volume is the target of the data in the replica set and cannot be accessed by the host. For disaster recovery purposes, if the primary volume is offline, the Standby volume is specified as the primary volume. The backup volume exists in the backup virtual disk of the backup system.

• Copy a snapshot. A special type of snapshot that retains the data status of the master volume of the replica set when a snapshot is created. After the master volume data is copied to the backup volume, the replication process creates a copy snapshot on both the master and backup systems. Although the copied snapshots are included in the maximum number of volumes of the system, they cannot be mapped and are not included in the license limit. You can copy snapshots to export them as normal licensed snapshots.

• Copy an image. Concepts of copying snapshots with the same image ID in the master and backup systems. The following synchronized snapshots contain the same data and can be used for disaster recovery.

Overview

As described in the simplified overview of the remote replication process, it can be configured to provide single-point real-time replication of volume data or regular incremental update replication. The regular update process includes multiple steps. In each step, a matched snapshot is created: In the master system, a copy snapshot of the current data of the master volume is created, and then the snapshot is used to add the new volume) data is copied from the primary volume to the standby volume. Next, a matching snapshot is created for the updated backup volume in the standby system. This creates replication synchronization points for matching snapshots. These synchronization points are used to continue the replication process. The following is a step-by-step example of the remote replication process.

It shows that two hosts use three replication groups:

• Hosts in New York map to the fiscal volume and update the fiscal volume. This volume is copied to the host in Munich.

• Hosts in Munich are mapped to the sales and engineering volumes and updated. The sales volume is copied from System 2 to system 3 of the Munich data center. The engineering volume is copied from system 3 in Munich to system 1 in New York.

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131228/00511I3a-0.jpg "title =" 1.jpg"/>

Remote replication uses the snapshot function to track the data to be copied and identify the data differences updated on the master volume, thus minimizing the amount of data to be transferred. The Snapshot created during the remote replication process is a special form of copying a snapshot, which does not count against the snapshot license restrictions.

In order to perform the replication, the master volume snapshot is captured to create the time point image of the data. Then, copy the data represented by the snapshot by using transmission media such as TCP/IP (iSCSI) or fibre channel) to copy the time point image to the backup volume. The first copy copies all data from the primary volume to the standby volume. The subsequent copies use sparse snapshots.

Both the master and backup volumes are retained. When the matching snapshot pairs of the two volumes are retained at the same time, the matching snapshot is called the replication synchronization point. These two snapshots are used together on each volume) as the synchronization reference point to minimize the amount of data to be transmitted. Two snapshots in a synchronization point are specified with the same image ID, which uniquely identifies the data in these snapshots from the same time point image and is the same in each block.

When a snapshot is created from a standard snapshot, and a standard snapshot still exists, the total data of the snapshot is expressed as zero bytes. The snapshot data is still associated with the standard quick camera and does not have the unique data associated with the copy quick camera. If a standard snapshot is deleted, its data will be retained by the snapshot associated with the copy snapshot), and the change in the copy snapshot size can reflect the size of the deleted snapshot.

One additional advantage of copying using snapshots is that these snapshots can be retained and can be restored in the future when non-hardware faults such as virus attacks occur. Because the copy source is a snapshot, this task does not copy any write operations performed on the master volume after the snapshot is captured. This allows you to better control the content contained in each copy image.

Snapshot operations require a large number of I/O operations. After a snapshot is created, each write to the unique position in the primary volume will result in internal read and write operations to keep the snapshot data. If you want to create a snapshot, create a volume copy, or copy a volume in a virtual disk, make sure that the master volume, snapshot pool, or the number of the two in the Virtual Disk cannot exceed four. For example, two primary volumes and two snapshot pools; three primary volumes and one snapshot pool; four primary volumes and zero snapshot pools.

Copy operation

650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131228/00511LS8-1.jpg "title =" 2.jpg"/>

1. Capture the initial snapshot And Start copying.

2. Initial replication includes full data replication.

3. Capture the second snapshot and request copying. This can be performed during initial replication.

4. Snapshots captured on the backup volume. This is the first replication synchronization point.

5. After the initial copy is completed, the second copy is automatically started. Only data changed since snapshot 1 is copied.

6. Snapshots taken on the backup volume. This is the second synchronization point.

7. Other snapshots can be captured and copied on the master volume during the replication process. These snapshots will be queued up until the previous copy is complete. The snapshots are not synchronized until the replication is complete.

Displays initial, incremental, and queuing replication:

Initial Replication: After the first copy is started, a snapshot of the master volume is taken and each data block is copied to the backup volume. After the replication is complete, the first snapshot is created on the secondary volume to create the first synchronization point. This synchronization point can be used to determine the incremental data from the synchronization point to the snapshot later. Operation 1-4 is the initial copy.

Incremental replication: The incremental data is a "list" of 64 KB blocks. The list is different between the last copied snapshot and the next snapshot to be copied. Then, the incremental data is copied from the master volume's copy snapshot to the backup volume. Once the initial replication is completed, as long as the synchronization point is maintained, all future replication of the replication group will be incremental replication. Operation 5 is incremental replication.

Queuing Replication: You can start a new copy when copying another snapshot. This allows you to take snapshots at specified intervals during other replication processes. Note: During the Process of another replication to the same backup volume, the replication started will wait in queue and data transmission will not start until the previous replication is complete. In operation 3, snapshot 2 is waiting in queue during snapshot 1 copying; in operation 7, snapshot 3 is waiting in queue during snapshot 2 copying. In the event of a network failure, you can manually pause ongoing replication or automatically execute it. If you want to continue the replication, you must manually restore it; or if you want to cancel the replication, you can also stop it.

Perform initial replication locally or remotely

When you set the volume replication, You need to specify the local master) system or remote backup) the backup volume in the system's virtual disk.

• If the initial replication speed is extremely important, specify the Virtual Disks owned by the same controller as the Virtual Disks of the primary volume in the local system. After setting replication, you can perform initial replication and then move the virtual disks that contain the backup volume and its snapshot pool to the remote system. Moving a virtual disk includes detaching a backup volume and stopping its virtual disk, removing the disk or chassis from the virtual disk, transporting the disk or chassis to a remote location, and inserting the disk or chassis into a remote system, and use SMU to restart the virtual disk and re-attach the backup volume. If the snapshot pool of the backup volume is in a different virtual disk, you must stop, move, and restart the virtual disk.

• If the ease of setting is the most important, specify a virtual disk owned by the Controller in the remote system. After setting replication, you can start copying. In either case, you must specify the FC or iSCSI Link Used for replication between the primary and standby systems and cannot change this setting when the replica set exists. You can only select a virtual disk with enough available space for replication.

Select a virtual disk to include the standard secondary volume

You can select an existing virtual disk to create a secondary volume when setting the volume of the primary volume that will become the replica set.

The Virtual Disk selection option only lists the volumes with sufficient available space for replication and does not contain conflicting names (rprimary-volume-name) or have conflicting names (sprprimary-volume-name) the virtual disk of the snapshot pool. The system calculates the space required by the secondary volume retention and its snapshot pool in the following ways:

• The Snapshot pool size is 20% or 6 GB of the master volume, which is larger than that of others ).

• The reserved size is calculated as follows:

• If both the master volume and the snapshot pool are smaller than 500 GB, the reserved size is the same as that of the master volume.

• If the primary volume is larger than 500 GB, the reserved size is the maximum value, that is, 500 GB.

• If the snapshot pool is larger than 500 GB, the retained size is as large as that of the snapshot pool.

• The calculation of the required space in the virtual disk is as follows:

• If the Combination Size of the master volume and the retained volume is smaller than the Combination Size of the master volume and the snapshot pool, the required space is the Combination Size of the master volume and the snapshot pool.

• If the size of the combination of the master volume and the retained volume is greater than the size of the combination of the master volume and the snapshot pool, the space required is the size of the combination of the master volume and the retained volume. The following table shows the available space of a virtual disk. If the volume size you want to copy is not displayed, you can use the above calculation method to determine the available space required for the secondary virtual disk.

Remote replication disaster recovery

Replication can continue when the following system faults occur, for example:

• Temporary communication faults. Remote replication retries the replication operation based on the policy configured by the user.

• Controller failure. In the dual-controller system, a Failover will take over the replication process until the Controller recovers.

• Disk or power failure.

If the master volume becomes inaccessible due to a disaster, you can set the backup volume to the master volume so that the volume can be mapped to the host. Disaster recovery requires user intervention because decisions must be made based on the data content of the copied volume and its snapshots.

1. Synchronize the backup volume with the copy snapshot, preferably with the copy synchronization point. Any data written to the master volume since the last copy is not available.

2. After synchronization, set the backup volume to be the new master volume.

3. Map the new primary volume to the host. The primary volume should be the same as the original primary volume.

   650) this. width = 650; "src =" http://www.bkjia.com/uploads/allimg/131228/00511Mb4-2.jpg "title =" 3.jpg"/>

If the original master volume becomes accessible, you can set it as the master volume again in the following procedure overview.

1. Snapshot of the original primary volume. This will retain the current data status of the volume for future comparison with the new primary volume.

2. Set the original master volume to a backup volume.

3. Copy all data written to the new primary volume to the original primary volume, which is now a backup volume ). This can be done through one or more copies. During the final replication, stop accessing the host of the master volume to ensure that all data has been transferred to the backup volume.

4. Set the original primary volume of the backup volume to a new primary volume.

5. Now, you can install, provide, and map the snapshot captured in step 1 and compare it with the new master volume to determine all the data differences, and try to recover any data that may be lost from the snapshot. For example, you can use the host file system tool to find any files modified since a specific time, or export any differential records from the snapshot, and then re-enter the current database.