40, "Huawei Hcie-storage"--globalcache

Source: Internet
Author: User

------------------------------------Important Note------------------------------------

The following sections are part of the network, some of which are from the Huawei store official materials

For specific materials, please visit Huawei Storage website for download

Network reference content can not be found original, if there is infringement please notify

------------------------------------Important Note------------------------------------


The global cache makes the memory space on all storage servers logically a pool of unified memory resources, slowing down the presence of data on any one storage server, and can be hit in the global cache when any subsequent storage server receives access to the data request. At the same time, all user data in the whole cluster system cache only one copy and do not cache the checksum data, can effectively play the impact of cache on system performance to improve the overall performance of the system.

①global cache and data read

① host E sends a data read request to node C's client.

② Node C's client requests a stripe resource read lock to the distributed lock server.

③ The system checks if the cache of the read data is in the global cache and on which node it is cached, and displays the file's sub-resources in the cache on Node A.

④ node C receives data from the global cache on Node A nodes and returns (if not in the global cache, after all stripe data of the stripe data is read directly from each node, then the striped data is constructed and then returned).

Distributed lock Management (DLM) is the basis for ensuring that the global cache is running efficiently, ensuring global cache sharing and consistency. Distributed lock management is responsible for creating a distributed lock management data structure that includes other related content such as shared resource lock requests, memory for storage shared resources, and lock types. A shared resource always exists whenever a process has a lock request on it, and the fabric lock manager can delete the resource if no process has a lock request for that resource. If the process exits unexpectedly. The lock associated with the resource is also exited abnormally. The lock associated with the resource is also released abnormally.

②global cache and data write


① host E sends a data write request to the client on node C.

② Node C's client requests a stripe resource write lock to the distributed lock server.

③ when node C's client application lock succeeds, node C receives the user data in the global cache on this node.

④ The client of Node C will fragment the stripe data according to the specified protection level, and the source of all shards is the data is computed by erasure code to generate the check data slice, and finally the data slice including the calibration data slice is written to the memory of the corresponding node, The write operation succeeds when the power-on memory of a node is written.

40, "Huawei Hcie-storage"--globalcache

Contact Us

The content source of this page is from Internet, which doesn't represent Alibaba Cloud's opinion; products and services mentioned on that page don't have any relationship with Alibaba Cloud. If the content of the page makes you feel confusing, please write us an email, we will handle the problem within 5 days after receiving your email.

If you find any instances of plagiarism from the community, please send an email to: info-contact@alibabacloud.com and provide relevant evidence. A staff member will contact you within 5 working days.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

  • Sales Support

    1 on 1 presale consultation

  • After-Sales Support

    24/7 Technical Support 6 Free Tickets per Quarter Faster Response

  • Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.