This is the 13th day of my participation in the August More Text Challenge. For details, see “August More Text Challenge”.
LUN snapshot (HyperSnap For Block)
HyperSnap For Block (LUN snapshot) can generate a consistent image of the source LUN at a certain point in time and quickly obtain a data copy consistent with that of the source LUN without interrupting normal services. The copy is available immediately after it is generated, and the read and write operations on the copy do not affect the source data. Therefore, through the snapshot technology can be solved such as online backup, data analysis, application testing and other problems. A LUN snapshot is implemented using the mapping table and copy-on-write technology.
Features of LUN snapshot technology
- Zero backup window
Traditional backup may degrade the performance of application hosts and even interrupt user services. Therefore, traditional backup must be performed when the application is down or the service volume is low. Backup window refers to the amount of time an application can tolerate to complete a data backup job. In fact, it is the amount of downtime an application can tolerate. Snapshots can be used to perform backup services online. The backup window is almost zero, and services do not need to be stopped.
- Save hard disk space
When a snapshot is used to obtain a consistent copy of the source LUN at the snapshot point in time, COW volume is used to store the data updated for the first time on the source LUN after the snapshot point in time. The size of COW volume has nothing to do with the source LUN, but depends on the change of the data on the source LUN after the snapshot point in time. When the amount of data on the source LUN does not change much, the snapshot provides a consistent copy of the source LUN with a small amount of disk space, which can be used by other testing services, saving disk space greatly.
- Fast data recovery
In traditional offline backup, the backup data cannot be read directly online, and the available copy of the original data at the backup point in time can be obtained only after a long data recovery process. In this way, the data can be restored. OceanStor snapshots can obtain the original data at the snapshot point in time by directly reading the snapshot volume. When the data on the source LUN is damaged, the data at the snapshot point in time can be directly recovered from the snapshot volume, facilitating data rollback.
- Activating snapshot Consistency
In OLTP applications, snapshots must be created for multiple copies of source LUN data at the same point in time to keep the associated application data on different luns at the same point in time. If snapshots cannot be created at the same time, applications may not work properly during data recovery using snapshots. For example, when a database is used, management data, service data, and log information are usually distributed on different source luns. Therefore, snapshots must be taken for the three source luns at the same point in time to ensure that the data on the three source luns is recovered to the same point in time. Otherwise, the data of the three parts can not be recovered to the same time point and the data correlation is lost, and the data recovery also loses its significance. The snapshot consistency activation of OceanStor storage array solves this problem by freezing the I/ OS that are being processed by multiple source luns in a consistency group at the snapshot point and obtaining the same snapshots of these source luns at the same time.
- Continuous data protection
OceanStor storage arrays can create snapshots for the same source LUN at multiple points in time. In combination with BCManager, snapshots can be created and deleted periodically at minute intervals. In BCManager, you can also set a policy for automatically activating snapshots or stopping snapshots periodically. This is a convenient and low-cost approximation of continuous data protection when multiple point-in-time snapshots are cyclically driven along the time axis to automate operations.
- The snapshot copy
Snapshot replica is a technology that backs up data at the time the snapshot is activated. It does not contain private data of the snapshot after the source snapshot is activated. The snapshot copy and source snapshot share the COW volume space of the source LUN, but the private space is completely independent. The snapshot copy is a writable snapshot that is completely independent of the source snapshot. The process for reading and writing a snapshot copy is the same as that for a snapshot. The snapshot copy technology allows you to obtain multiple data copies of the same snapshot. Multiple snapshot replicas can be created for different data purposes.
FS snapshot (HyperSnap For File)
HyperSnap For File generates snapshots of a source File system at a certain point in time and quickly obtains a data copy consistent with that of the source File system without interrupting services. The copy is available immediately after it is generated, and data in the source file system is not affected by read/write operations on the copy data. Therefore, the file system snapshot technology can solve problems such as online backup, data analysis, application testing and so on. You can use a file system snapshot in various ways. For example, they can be used for:
1. Create a snapshot for the file system and back up the snapshot data to tapes.
2. After a file system snapshot is created, end users can recover their own files from the snapshot if the file system snapshot is accidentally deleted or damaged.
3. Features such as remote replication and HyperVault require file system snapshots. Snapshot data can be replicated or backed up to a remote end.
File system snapshot is implemented using the Redirect On Write (ROW) file system technology. When new or modified data is written to a file system, the new data does not overwrite the old data. Instead, new space is allocated to the storage media for data writing. This ensures high data reliability and file system scalability. File system snapshots based on ROW technology that can be created quickly (in seconds) and do not take up extra disk space unless the original file is deleted or changed.
FS snapshot technology features
-
Zero Backup window A backup window is the amount of time an application can tolerate to complete data backup. In fact, it is the amount of downtime an application can tolerate. Traditional backup may degrade file system performance and even interrupt user services. Therefore, traditional backup jobs must be performed when applications are down or the traffic is low. When file system snapshots are used for backup services, the backup service can be performed online, the backup window is almost zero, and services do not stop.
-
Second level snapshot
Creating a file system snapshot is a copy of the roots of a tree. It takes a short time to create a second snapshot.
- Low performance loss
Creating a snapshot for a file system is simple, has little data on the footwall, and has little impact on system performance. After a snapshot is created, the I/O process of the file system only needs to check whether the data is protected by the snapshot before the data space is released, and record the space of the data blocks that are protected by the snapshot. The impact on the file system performance is negligible. After the snapshot is deleted, background data reclamation compets with file system services for SOME CPU and memory resources, but the performance loss is also at a low level.
- Save disk space
When a file system snapshot is used to obtain a consistent copy of the source file system at the snapshot point in time, the space exclusively occupied by the snapshot is determined by the data change of the source file system after the snapshot point in time and never exceeds the size of the file system at the snapshot creation time. When the data amount of the source file system does not change much, the file system snapshot provides a consistent copy of the source file system with a small amount of storage space, saving disk space greatly. 5. Quick access to snapshot data The file system snapshot is displayed as a single directory in the root directory of the file system. Users can access the directory corresponding to the snapshot to quickly access the snapshot data. In scenarios where snapshot rollback is not required, data at the snapshot point in time can be easily accessed and data can be recovered by copying files or directories when file data in the current file system is damaged. You can access a CIFS shared file system from a Windows client and restore a file or directory to the snapshot at a point in time. Simply right-click the directory or file to be restored, select the previous version, and you can see all the points in time that contain the snapshot of the file or directory. You can select one of the points in time to restore the data.
- The file system was quickly rolled back
In traditional offline backup, the backup data cannot be directly read online, and the available copy of the original data at the backup point in time can be obtained only after a long period of data recovery, thus realizing data restoration. For the unified file system snapshot, the root of the file system can be directly replaced with the root of the specified snapshot and the cached data can be cleared, so that the file system can be quickly rolled back to the specified snapshot point in time. You need to use the rollback command with caution, because after the file system rollback is complete, snapshots after the rollback time point are automatically deleted.
- Periodic snapshot provides continuous data protection
File system snapshot You can configure a policy to automatically create a snapshot at a specified time or at a specified time interval. The maximum number of periodic snapshots supported by a file system depends on the product model. If the maximum number of periodic snapshots is exceeded, the earliest snapshots at a point in time are automatically deleted without user intervention.
The file system also supports users to delete snapshots created periodically. This provides a convenient and low-cost approach to continuous data protection by taking snapshots of multiple points in time moving forward through the timeline. Continuous data protection (CDP) cannot be implemented by using snapshots. The minimum time interval between two snapshot points determines the granularity of Continuous data protection.
