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 File System Feature Guide 12

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).
Performance Tuning

Performance Tuning

This section provides suggestions on the methods to improve remote replication performance.

The impact on the replication performance is measured by the following:
  • The amount of data per unit time that is replicated from the primary end to the secondary end, namely, the replication bandwidth.
  • The impact of a replication task on the access performance of the front-end service.

InfoReplicator uses distributed links to balance loads across all the CPU, memory, and network resources, achieving the largest amount of data transferred per unit time. Also, this feature allows you to limit the link bandwidth and transmission rate to minimize the impact of remote replication on system performance.

If you want to improve the performance for some specific scenarios, take suggestions on InfoReplicator provided in this section for reference.
  • Adjusting system resources
    • Configure high-performance storage nodes for the storage systems. High-performance storage nodes are configured with better computing resources such as CPU and memory and therefore provide higher read and write performance.

      Disks: SSD > SAS > SATA

    • Improve the bandwidth of the front-end service network. Use a front-end service network that delivers a higher bandwidth to eliminate the replication performance bottleneck caused by the replication network.

      Under the same conditions, the bandwidth of a front-end service network is ranked as follows: IB network > 10GE network > GE network.

    • Add more nodes to the primary and secondary storage systems and ensure that the two storage systems have the same number of available storage nodes. Adding nodes to the storage systems enables more nodes to participate in replication and increases the CPU, memory, disk and network resources, allowing more data to be replicated per time unit. The same number of available storage nodes in the primary and secondary storage systems ensures access load balancing across nodes, eliminating the node performance bottleneck.
  • Adjusting the configuration of InfoReplicator
    • Do not perform full synchronization (except the initial synchronization). Full synchronization requires a large amount of systems resources and prolongs the replication time.
    • Increase the transmission rate of a pair. Normally, the priority of a replication task is set to Medium. Under this configuration, the resources allocated to remote replication decrease as multiple background tasks run in parallel and therefore the replication time becomes long. Increasing the transmission rate of a pair means a higher execution priority for a replication task. Resources will be preferentially allocated to a replication task of a high priority, improving the replication performance.

      However, the increase of transmission rate exerts a greater impact on the front-end service performance.

    • Increase the number of nodes in the replication zone. The increasing number of nodes in the replication zone means that a larger number of nodes will participate in replication.
    • Set a bandwidth limit for the replication channel. The bandwidth limit of the replication channel can restrict the bandwidth of nodes at the originating end of the replication channel, preventing a replication task from affecting other services by occupying a high bandwidth. As long as the system performance meets your requirement, you can shorten the replication time by canceling the bandwidth limit or setting a lower one to increase the data replication capacity of the nodes at the originating end of the replication channel.
  • Other adjustments
    • Set an appropriate directory redundancy ratio and strip size. Under the same conditions, the read and write performance is ranked as follows:

      Redundancy ratio: N+1 > N+2 = N+2:1 > N+3 = N+3:1 > N+4

      Strip size: 512KB > 256 KB > 128 KB > 32 KB > 16 KB
      • If the average file size is equal to or smaller than 64 KB, set the stripe size to 16 KB.
      • If the average file size is larger than 64 KB and equal to or smaller than 256 KB, set the stripe size to 32 KB.
      • If the average file size is larger than 256 KB and equal to or smaller than 2048 KB, set the stripe size to 128 KB.
      • If the average file size is larger than 2048 KB and equal to or smaller than 4096 KB, set the stripe size to 256 KB.
      • If the average file size is larger than 4096 KB, set the strip size to 512 KB.
      • If the average file size is larger than 8192 KB, set the strip size to 1024 KB.
    • Do not perform remote replication and NDMP backup in parallel because the two types of tasks will affect each other.
Translation
Download
Updated: 2019-06-27

Document ID: EDOC1000122519

Views: 69735

Downloads: 145

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