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

Reminder

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

upgrade

OceanStor 9000 V300R006C00 Object Storage Service (Compatible with OpenStack Swift APIs) Administrator Guide 07

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).
Data Reconstruction Mechanism for Disk or Node Failures

Data Reconstruction Mechanism for Disk or Node Failures

The OceanStor 9000 allows users to manually start data reconstruction tasks or automatically reconstructs data in the background in the specified period based on the preset policy.

Data Reconstruction and Trigger Conditions

Data reconstruction is a function provided by the OceanStor 9000 to retrieve data as long as the number of failed disks or nodes does not exceed the threshold specified by the data protection level or when a hardware fault or high service pressure causes a data write failure.

NOTE:

Before retrieving data, the OceanStor 9000 uses an internal algorithm to ensure data accessibility.

If the number of failed disks or nodes exceeds the threshold of specified by the data protection level, a data reconstruction task can still be executed. In this case, the OceanStor 9000 employs the best-effort policy to retrieve valid data as much as possible.

Table 9-1 describes the trigger conditions of data reconstruction.

Table 9-1  Trigger conditions of data reconstruction

Failed Object

Trigger Condition

Reconstruction Method

Automatic Reconstruct or Not

Disk

A disk is offline for not more than 10 minutes and goes offline not more than four times within 10 minutes.

Data changed when the disk is offline is updated to the disk after it is reconstructed. This method is available only when data on the disk is intact before the disk goes offline.

Yes

  • A disk is offline for more than 10 minutes.
  • A disk goes offline five times within 10 minutes.
  • A disk fails and cannot be brought online.

The system uses calculations to reconstruct data on the failed disk and writes the reconstructed data onto another normal disk.

Yes

Node

A node is taken offline and brought online again.

Data changed when the node is offline is updated to the node after it is reconstructed. This method is available only when data on the node is intact before the node goes offline.

Yes

A node goes offline, and an administrator manually triggers data reconstruction.

The system uses calculations to reconstruct data on the failed node and writes the reconstructed data onto another normal node.

No

-

A hardware fault or high service pressure causes a data write failure.

The system uses calculations to synchronize data failed to be written.

Yes

Automatic Data Reconstruction Policy

Before performing data reconstruction tasks, plan automatic data reconstruction policies. The OceanStor 9000 supports a maximum of five policies. An automatic data reconstruction policy specifies the priority of a data reconstruction operation in a specific period. Periods specified in the five policies cannot overlap. Possible values of the policy priority are as follows:

  • High (default value): During the specified period, data reconstruction causes system performance to deteriorate by not more than 50%.

  • Medium: During the specified period, data reconstruction causes system performance to deteriorate by not more than 30%.

  • Low: During the specified period, data reconstruction causes system performance to deteriorate by not more than 10%.

  • Stop: The data reconstruction task is stopped.

  • Full speed: Flow control is not performed for the data reconstruction task.

When data protection is about to become invalid, the policy priority of the corresponding data reconstruction task automatically becomes High. For example, when two disks fail in a logical group whose protection level is 4+2, the policy priority of the data reconstruction task of the logical group automatically becomes High.

Working Principles of Quick Data Reconstruction

As shown in Figure 9-1, based on the specific load balancing policy and data protection level, the OceanStor 9000 internally divides storage resources into multiple logical groups, and each disk may belong to multiple logical groups. A logical group is an independent unit for data protection and data reconstruction. Each logical group is assigned a data protection level (for example, 4+2).

If a disk failure triggers data reconstruction, all the logical groups to which the failed disk belongs will participate in the data reconstruction at the same time. In this way, the reconstruction speed is faster than that of a traditional RAID group.

If a node failure triggers data reconstruction, all the logical groups to which the node's disks belong will participate in the data reconstruction at the same time. The maximum reconstruction speed is 1 TB/hour.

Figure 9-1  Schematic diagram of data storage distribution

Translation
Download
Updated: 2019-04-28

Document ID: EDOC1000122524

Views: 10650

Downloads: 86

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