No relevant resource is found in the selected language.

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>Search


To have a better experience, please upgrade your IE browser.


FusionStorage V100R006C30 Block Storage Service Disaster Recovery Feature Guide 03

Rate and give feedback :
Huawei uses machine translation combined with human proofreading to translate this document to different languages in order to help you better understand the content of this document. Note: Even the most advanced machine translation cannot match the quality of professional translators. Huawei shall not bear any responsibility for translation accuracy and it is recommended that you refer to the English document (a link for which has been provided).
HyperMetro I/O Processing Mechanism

HyperMetro I/O Processing Mechanism

HyperMetro implements the active-active relationship between the data in two data centers through the dual-write and Data Change Log (DCL) mechanisms. In addition, HyperMetro enables the storage systems in the two data centers to concurrently provide data read and write capabilities for hosts.

Write I/O Process

HyperMetro uses the efficient optimistic locking and dual-write mechanisms to ensure data consistency between two storage devices.
  • Data changes can be synchronized using the dual-write and DCL technologies while services are running. The dual-write technology enables I/O requests of application servers to be delivered to both the local and remote storage pools. In this way, the consistency of changed data can be ensured between the local and remote storage pools. If a DC fails, the DCL records data changes in running DCs, and traces changed data after the DC recovers, ensuring that all changed data is synchronized to storage devices in the DC.
  • Two storage systems enabled with HyperMetro can process host I/O requests concurrently. When multiple hosts access a storage address at the same time, an access conflict occurs. To prevent such conflicts, an optimistic lock mechanism is required. All dual-write operations are performed at both the local and remote sites, and a local lock is added at both sites. When the lock detects that some I/Os have address conflicts, these I/Os are forwarded to the other site for processing.
Figure 1-1 shows the write I/O process when an I/O request is delivered from an application host and causes data changes.

In the following figure, the write I/O accesses the local storage pool, and the local storage pool writes data into the remote storage pool for dual-write purposes.

Figure 1-1  Write I/O process

  1. An application host delivers a write I/O request to the DR cluster in HyperMetro service.
  2. A log is recorded.
  3. Dual-write is performed. That is, the HyperMetro service simultaneously writes the data into the local and remote storage pools.
  4. The local and remote storage pools return the write I/O result.
  5. The write I/O result is returned to the application host only after the write I/O results are received from both the local and remote storage pools.
  6. The storage system determines whether dual-write is successful.
    • If the data is written into both storage systems, the log is deleted.
    • If dual-write fails on one site, the log is converted into a DCL to record the differential data between the local and remote storage pools.

      If the data fails to be written into one storage pool, the HyperMetro service is paused and the storage pool in each data center sends an arbitration request to the quorum server. The storage pool that wins the arbitration continues providing services and the one that fails stops. Data is synchronized in the background based on the recorded DCL. When data is consistent between the local and remote storage pools, the HyperMetro service recovers.

Read I/O Process

Data in the volumes of both storage pools are synchronized in real time and both are accessible to hosts.

Figure 1-2 shows the read I/O process.
Figure 1-2  Read I/O process

  1. A host sends a read request to the HyperMetro service of the DR cluster.
    When links between storage systems in two data centers are faulty, the HyperMetro service responds to the host request based on arbitration results of the quorum server.
  2. The DR cluster enables the local storage system to respond to the read I/O request made by the host.
  3. If the local storage system is working properly, it returns data to the HyperMetro service.
  4. The read I/O request sent by the host is successfully processed.
Updated: 2019-01-17

Document ID: EDOC1100044928

Views: 17299

Downloads: 33

Average rating:
This Document Applies to these Products
Related Documents
Related Version
Previous Next