CN-121255114-B - Storage space scrubbing optimization method, system, device, chip, network interface card, apparatus, medium and program product

CN121255114BCN 121255114 BCN121255114 BCN 121255114BCN-121255114-B

Abstract

The application relates to a storage space scrubbing optimization method, a system, a device, a chip, a network interface card, equipment, a medium and a program product. The method comprises the steps of receiving information of a first storage area of a service application, carrying out hot zone marking on the first storage area based on the information of the first storage area, and skipping the first storage area with the hot zone marking or increasing a scanning period of scanning scrubbing of the first storage area with the hot zone marking under the condition of scanning scrubbing of a storage space. The method can effectively avoid the scrubbing action from interfering with the service flow, and reduce the performance jitter.

Inventors

LIU JINGTAO
CHEN ANQING
He Chaoqian

Assignees

杭州云豹创芯智能科技有限公司
深圳云豹智能股份有限公司

Dates

Publication Date: 20260512
Application Date: 20251208

Claims (14)

1. A method for optimizing storage space scrubbing, the method comprising: Receiving information of a first storage area of a service application, wherein the information of the first storage area is an address range corresponding to a service process; based on the information of the first storage area, carrying out hot zone marking on the first storage area; In the case of scanning scrubbing a storage space, skipping the first storage area where a hot zone mark exists or increasing a scanning period of scanning scrubbing of the first storage area where the hot zone mark exists.
2. The method according to claim 1, wherein the method further comprises: Receiving information of a second storage area corresponding to business exit, wherein the second storage area is at least one of the first storage areas; And deleting the hot zone mark of the second storage area based on the information of the second storage area.
3. The method according to claim 1, wherein the method further comprises: collecting load information corresponding to the storage space in a polling mode; and adjusting a scanning period for scanning and scrubbing the storage space based on the load information.
4. The method of claim 3, wherein the adjusting a scan period for scanning scrubbing the memory space based on the load information comprises: Increasing a scanning period for scanning and scrubbing a third storage area under the condition that the load rate of the storage space is larger than or equal to a threshold value based on the load information, wherein the third storage area is a storage area without a hot zone mark in the storage space; And setting the scanning period of scanning scrubbing on the third storage area as the scanning period of default scanning scrubbing under the condition that the loading rate of the storage space is determined to be smaller than a threshold value based on the loading information.
5. The method according to claim 4, wherein the method further comprises: and under the condition that the correctable error of the fourth storage area is received, adjusting the scanning period of the scanning scrubbing of the fourth storage area, wherein the correctable error of the fourth storage area is reported when verification failure occurs in the scanning scrubbing process, and the fourth storage area comprises the first storage area and the third storage area.
6. The method of claim 5, wherein said adjusting the scan period of the scan scrub of the fourth storage area comprises: When a hot zone mark exists in the fourth storage area and the physical address range corresponding to the fourth storage area belongs to a completely skipped address range, adjusting the scanning period of the scanning scrubbing of the fourth storage area to be the scanning period of the scanning scrubbing of the storage area with the hot zone mark; Reducing the scanning period of the scanning scrubbing of the fourth storage area under the condition that the hot zone mark exists in the fourth storage area and the physical address range corresponding to the fourth storage area does not belong to the completely skipped address range; and reducing the scanning period of the scanning scrubbing of the fourth storage area in the condition that the fourth storage area is not provided with a hot zone mark.
7. The method according to any one of claims 1 to 6, further comprising: generating scrubbing operations of the storage controller based on information of the storage area to be scanned and scrubbed; obtaining a scrubbing interval of the storage controller based on a scanning period of scanning scrubbing of the storage area to be scanned and scrubbed; and scanning and scrubbing the storage area to be scanned and scrubbed based on the scrubbing interval and the scrubbing operation.
8. A storage space scrubbing system, which is characterized by comprising a data processing unit and a storage controller corresponding to a storage space; the data processing unit is configured to perform the storage space scrubbing optimization method of any one of claims 1 to 7; The storage controller is used for scanning and scrubbing the storage space based on the control of the data processing unit; Or alternatively The system comprises a data processing unit, a storage controller corresponding to the storage space and a scheduling engine; The data processing unit is used for sending information of a first storage area of the service application to the dispatching engine; the scheduler engine configured to perform the storage space scrubbing optimization method of any one of claims 1 to 7; the storage controller is used for scanning and scrubbing the storage space based on the control of the data processing unit.
9. A storage space scrubbing optimizing apparatus, said apparatus comprising: the receiving module is used for receiving information of a first storage area of the service application, wherein the information of the first storage area is an address range corresponding to the service process; The marking module is used for carrying out hot zone marking on the first storage area based on the information of the first storage area; And the scanning period adjustment module is used for skipping the first storage area with the hot zone mark or increasing the scanning period of the scanning scrubbing of the first storage area with the hot zone mark under the condition of scanning scrubbing of the storage space.
10. A chip comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 7 when the computer program is executed.
11. A network interface card comprising a chip and a plurality of interfaces as recited in claim 10, The chip processes data or communicates with the outside through the interface.
12. A computer device comprising the network interface card of claim 11 for processing data or communicating externally.
13. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 7.
14. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 7.

Description

Storage space scrubbing optimization method, system, device, chip, network interface card, apparatus, medium and program product Technical Field The present application relates to the field of computer technology, and in particular, to a method, system, device, chip, network interface card, device, medium, and program product for optimizing scrubbing of a storage space. Background DPU (Data Processing Unit) is a data-centric dedicated processor that provides network, storage, security, management, etc., data center infrastructure virtualization services around the data processing. Performance improvement and time delay reduction are realized through storage/network/safety hardware unloading. The DPU acts as a "third dominant chip" for the data center following CPU (Central Processing Unit) and GPU (Graphics Processing Unit), and its stable operation is critical to the data center performance, safety, and energy efficiency. In the conventional technology, a periodic full memory scan may cause bus bandwidth occupation, for example, a scanner needs to periodically scan a memory (e.g., every 24 hours full disk scan), and read data triggers ECC (Error Correction/Correction Code) verification, where a scanner (Scrubber) of a DDR controller is a hardware or software mechanism that actively detects and repairs soft errors of a memory (e.g., single bit flip caused by cosmic rays or electromagnetic interference), and can prevent system crashes or data damages caused by Error accumulation, which is one of key technologies for guaranteeing the reliability of the memory. However, the process of scanning the memory occupies the memory bus bandwidth, and the scanning bandwidth overhead can reach 1% -5% in the DDR4 and DDR5 high-frequency interfaces. High frequency scanning, while improving reliability, can squeeze traffic application bandwidth, especially in high throughput scenarios (e.g., AI training, database) where performance jitter may be incurred. Disclosure of Invention In view of the foregoing, it is desirable to provide a method, system, device, chip, network interface card, apparatus, medium, and program product for optimizing scrubbing of a memory space that can effectively avoid the scrubbing behavior from interfering with traffic flows and reduce performance jitter. In a first aspect, the present application provides a method for optimizing scrubbing of a storage space, the method comprising: receiving information of a first storage area of a service application; based on the information of the first storage area, carrying out hot zone marking on the first storage area; In the case of scanning scrubbing a storage space, skipping the first storage area where a hot zone mark exists or increasing a scanning period of scanning scrubbing of the first storage area where the hot zone mark exists. In one embodiment, the method further comprises: Receiving information of a second storage area corresponding to business exit, wherein the second storage area is at least one of the first storage areas; And deleting the hot zone mark of the second storage area based on the information of the second storage area. In one embodiment, the method further comprises: collecting load information corresponding to the storage space in a polling mode; and adjusting a scanning period for scanning and scrubbing the storage space based on the load information. In one embodiment, the adjusting the scanning period of scanning and scrubbing the storage space based on the load information includes: Increasing a scanning period for scanning and scrubbing a third storage area under the condition that the load rate of the storage space is larger than or equal to a threshold value based on the load information, wherein the third storage area is a storage area without a hot zone mark in the storage space; And setting the scanning period of scanning scrubbing on the third storage area as the scanning period of default scanning scrubbing under the condition that the loading rate of the storage space is determined to be smaller than a threshold value based on the loading information. In one embodiment, the method further comprises: and under the condition that the correctable error of the fourth storage area is received, adjusting the scanning period of the scanning scrubbing of the fourth storage area, wherein the correctable error of the fourth storage area is reported when verification failure occurs in the scanning scrubbing process, and the fourth storage area comprises the first storage area and the third storage area. In one embodiment, the adjusting the scan period of the scan scrub of the fourth storage area includes: When a hot zone mark exists in the fourth storage area and the physical address range corresponding to the fourth storage area belongs to a completely skipped address range, adjusting the scanning period of the scanning scrubbing of the fourth storage area to be the scanning period of the scanning scrubbing of the storage ar