WO2024055543A1

WO2024055543A1 - Non-transparent bridge management method and apparatus, controller, and medium

Info

Publication number: WO2024055543A1
Application number: PCT/CN2023/080892
Authority: WO
Inventors: 钟戟; 王鲁泮
Original assignee: 苏州元脑智能科技有限公司
Priority date: 2022-09-13
Filing date: 2023-03-10
Publication date: 2024-03-21
Also published as: CN115202938B; CN115202938A

Abstract

The present application discloses a non-transparent bridge management method and apparatus, a controller, and a nonvolatile medium. The method comprises: partitioning data to be backed up and distributing channels for each region; by means of all the channels of each region, continuously sending data to a first non-transparent bridge connected to a main controller, so that the first non-transparent bridge separately performs continuity verification on the received data corresponding to each region; and acquiring a first verification result returned by the first non-transparent bridge, and if the first verification result is that a certain region does not pass the continuity verification, enumerating and verifying all the channels of the region. Therefore, a problematic channel connected to a non-transparent bridge can be accurately verified, the problem can be processed in a timely manner, the accuracy and efficiency of non-transparent bridge verification management can be improved, and thus the accuracy of data backup is ensured.

Description

A non-transparent bridge management method, device, controller and medium

Cross-references to related applications

This application requests the priority of the Chinese patent application submitted to the China Patent Office on September 13, 2022, with the application number 202211107200.8, and the application name is "A non-transparent bridge management method, device, controller and medium", and its entire content incorporated herein by reference.

Technical field

The present application relates to the field of storage technology, and in particular to a non-transparent bridge management method, device, controller and non-volatile readable storage medium.

Background technique

In the era of big data, higher requirements are placed on the reliability of storage arrays, especially the reliability and stability of data backup in storage systems. NTB (Non-Transparent Bridging, non-transparent bridge) is currently used in the storage system to realize data backup between dual controllers to ensure that the data processed on the dual controllers can be backed up quickly and accurately. Using this kind of mirror backup greatly improves the efficiency of data backup. improve the reliability of storage devices. However, in the process of applying a non-transparent bridge chip to achieve data backup, the controller is connected to the NTB chip for data transmission through the PCIE (Peripheral Component Interconnect Express, high-speed serial computer expansion bus standard) channel. During the PCIE serial data transmission process , it is difficult to verify and manage multiple channels connected by NTB. When a PCIE link is abnormal, it is difficult to locate and analyze the abnormal problem. Abnormal paths may cause errors in the controller's data backup process, eventually leading to data storage errors. Even serious problems such as disk loss; at the same time, with the development of current PCIE technology, the bandwidth required to transmit and back up data is gradually increasing, and as the functions of the storage motherboard become more complex, the length of the data transmission channel on the motherboard will also have a certain limit. Extension, the above series of physical factors bring greater difficulties to data transmission backup on the PCIE channel, and the difficulty of the NTB chip to verify the accuracy of PCIE channel data transmission is also further increased. Therefore, how to accurately and quickly perform NTB verification Management to ensure the accuracy of the data backup process is an urgent problem that needs to be solved.

Contents of the invention

In view of this, the purpose of this application is to provide a non-transparent bridge management method, device, controller and non-volatile readable storage medium, which can improve the accuracy and efficiency of NTB verification management, thereby ensuring the accuracy of data backup sex. The plan is as follows:

In the first aspect, this application discloses a non-transparent bridge management method, which is applied to the main controller in a dual controller, including: partitioning the data to be backed up and allocating channels to each zone; passing all channels in each zone to The first non-transparent bridge connected to the main controller continuously sends data, so that the first non-transparent bridge performs continuity check on the received data corresponding to each zone; obtains the first check returned by the first non-transparent bridge As a result, if the first verification result is a certain If the zone fails the continuity check, all channels in the zone will be enumerated and checked.

Optionally, partition the data to be backed up and allocate channels to each zone, including: horizontally partitioning the data to be backed up according to the data service object to obtain each zone; vertically partitioning the data to be backed up according to the importance of the data to obtain each vertical partition ;Assign channels to each zone based on the amount of data each zone has in each vertical partition.

Optionally, horizontally partition the data to be backed up based on the data service object to obtain each area, including: obtaining the data service object from the front-end server; horizontally partitioning the data to be backed up based on the data service object to obtain each area.

Optionally, it also includes: if the first verification result is that each zone passes the continuity check, start data backup, and back up the data to be backed up to the dual controller according to each zone and through the channel of each zone. backup controller.

Optionally, after partitioning the data to be backed up and allocating channels to each zone, it also includes: sending the partition and channel allocation to the backup controller, so that the backup controller can process the data through each zone based on the partition and channel allocation. All channels continuously send data to the second non-transparent bridge connected to the backup controller, so that the second non-transparent bridge performs continuity check on the received data corresponding to each zone, and returns on the second non-transparent bridge When the second verification result is that a certain area fails the continuity check, all channels in the area will be enumerated and verified.

Optionally, if the first verification result is that each area passes the continuity check, data backup is started, including: if the first verification result is that each area passes the continuity check, and the second verification The result is that each zone passes the continuity check, and data backup is started.

Optionally, perform enumeration and verification on all channels in this area, including: continuously sending data through the channel in this area to the first non-transparent bridge connected to the main controller, so that the first non-transparent bridge can respectively receive The data corresponding to each channel is checked for continuity.

Optionally, the data to be backed up is backed up to the backup controller in the dual controller according to each zone and through the channel of each zone, including: dividing the data to be backed up into batches, obtaining multiple batches of data, and converting the multiple batches into The data is backed up to the backup controller in the dual controller batch by batch according to each zone and through the channel of each zone; accordingly, the method also includes: after the backup of any batch of data is completed, extract the first feature of the batch of data and The first feature is transmitted to the first non-transparent bridge, so that the first non-transparent bridge compares the first feature and the second feature to see if they are consistent; wherein, the second feature is the batch of data received by the backup controller extracted by the backup controller. Second characteristic.

Optionally, the method further includes: sending an execution instruction to the backup controller, so that the backup controller extracts the second characteristic of the received batch data under the control of the execution instruction.

Optionally, the method further includes: obtaining the second characteristic sent by the backup controller; and transmitting the second characteristic to the first non-transparent bridge, so that the first non-transparent bridge compares whether the first characteristic and the second characteristic are consistent.

Optionally, it also includes: if the first feature and the second feature are inconsistent, determine the problem channel and assign the problem channel closure.

Optionally, determining the problem channel includes: continuously sending data through all channels in each zone to the first non-transparent bridge connected to the main controller, so that the first non-transparent bridge receives data corresponding to each zone. The data is checked for continuity; the check result returned by the first non-transparent bridge is obtained. If the check result is that a certain area fails the continuity check, all channels in the area are enumerated and checked, and the Channels that fail the verification are determined to be problem channels.

Optionally, extracting the first feature of the batch of data includes: determining a hash function; using the hash function to extract a set of hash values of the batch of data to obtain the first feature of the batch of data.

Optionally, determining the hash function includes: identifying the data type of the batch of data; determining the hash function based on the data type.

Optionally, use a hash function to extract a hash value set of the batch of data to obtain the first feature of the batch of data, including: for any data type in the batch of data, use the hash value corresponding to the data type. The column function extracts hash values; the hash values corresponding to different data types are merged into the hash value set of the batch of data to obtain the first feature of the batch of data.

Optionally, the data type includes text type, image type, and video type. Correspondingly, determining the hash function based on the data type includes: constructing the first hash function based on the division-leaving-remainder method; determining the hash function of the text type as A first hash function; a second hash function is constructed based on the squared median method; and the hash function of the image type and the video type is determined as the second hash function.

In the second aspect, this application discloses a non-transparent bridge management device, which is applied to the main controller in a dual controller, including: a data partition module, which is set to partition the backup data; a channel allocation module, which is set to Each zone is allocated a channel; the zone check module is configured to continuously send data through all channels of each zone to the first non-transparent bridge connected to the main controller, so that the first non-transparent bridge responds to each received The data corresponding to the area is checked for continuity; the check result acquisition module is set to obtain the first check result returned by the first non-transparent bridge; the channel check module is set to if the first check result is a certain If the zone fails the continuity check, all channels in the zone will be enumerated and checked.

In a third aspect, this application discloses a controller that includes a memory and a processor, wherein: the memory is configured to save a computer program; the processor is configured to execute the computer program to implement the aforementioned non-transparent bridge management method.

In a fourth aspect, the present application discloses a computer non-volatile readable storage medium configured to store a computer program, wherein the computer program implements the aforementioned non-transparent bridge management method when executed by a processor.

It can be seen that this application first partitions the backup data and allocates channels to each zone, and then continuously sends data to the first non-transparent bridge connected to the main controller through all channels in each zone, so that the first non-transparent bridge can respectively take over The data corresponding to each zone is checked for continuity, and then the first check result returned by the first non-transparent bridge is obtained. If the first check result is that a certain zone fails the continuity check, the zone is All channels are enumerated and verified. That is to say, before data backup, this application partitions and allocates channels for the backup data, and then verifies each area first. If a certain area fails the verification, then verifies each channel. In this way, it can accurately Verifying the problematic channels connected on the non-transparent bridge and handling the problems in a timely manner can improve the accuracy and efficiency of non-transparent bridge verification management, thereby ensuring the accuracy of data backup.

Description of drawings

In order to explain the embodiments of the present application or the technical solutions in the prior art more clearly, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are only This is an embodiment of the present application. For those of ordinary skill in the art, other drawings can be obtained based on the provided drawings without exerting creative efforts.

Figure 1 is a flow chart of a non-transparent bridge management method disclosed in an embodiment of the present application;

Figure 2 is a schematic structural diagram of an optional dual-controller storage system provided by an embodiment of the present application;

Figure 3 is a flow chart of an optional non-transparent bridge management method disclosed in the embodiment of the present application;

Figure 4 is an optional data backup flow chart disclosed in the embodiment of the present application;

Figure 5 is a flow chart of an optional partition enumeration scanning algorithm disclosed in the embodiment of the present application;

Figure 6 is a flow chart of an optional data comparison algorithm based on the Hash principle disclosed in the embodiment of the present application;

Figure 7 is a schematic structural diagram of a non-transparent bridge management device disclosed in an embodiment of the present application;

Figure 8 is a schematic structural diagram of a controller disclosed in the embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of this application.

Currently, in the process of applying non-transparent bridge chips to realize data backup, the controller is connected to the NTB chip through the PCIE channel for data transmission. During the PCIE serial data transmission process, it is difficult to verify and manage the multiple channels of the NTB connection. When a When the PCIE link is abnormal, it is difficult to locate and analyze the abnormal problem. The abnormal path may cause errors in the controller's data backup process, eventually leading to serious problems such as data storage errors and even disk loss. At the same time, with the current development of PCIE technology, With the development of the industry, the bandwidth required for data transmission and backup is gradually increasing. With the complexity of storage motherboard functions, the length of the data transmission channel on the motherboard will also be extended to a certain extent. The above series of physical factors have limited the data transmission on the PCIE channel. Backup brings greater difficulties, and it becomes more difficult for the NTB chip to verify the accuracy of PCIE channel data transmission. Therefore, how to accurately perform NTB verification management to ensure the accuracy of the data backup process is an issue that needs to be solved urgently. To this end, this application provides a non-transparent bridge management solution that can improve the accuracy of NTB verification management and thereby ensure the accuracy of data backup.

As shown in Figure 1, an embodiment of the present application discloses a non-transparent bridge management method, which is applied to the main controller in a dual controller, including:

Step S11: Partition the data to be backed up and allocate channels to each zone.

For example, see Figure 2, which is a schematic structural diagram of an optional dual-controller storage system provided by an embodiment of the present application. The dual controllers realize mutual backup of data through non-transparent bridges. One controller has 16 PCIE channels connected to the non-transparent bridge, and the other controller also has 16 PCIE channels connected to the non-transparent bridge. Two non-transparent bridges. Transparent bridges are interconnected, which allows both controllers to access each other's systems through non-transparent bridges, and data can be backed up each other. PCIE transmits data through serial transmission, so the start and end of data transmission between channels will not interfere with each other. Each PCIE channel from the two controllers to the NTB is independent of each other. This serial transmission method The data transmission rate is increased, but the independence of each channel increases the complexity of channel verification. The embodiment of the present application can accurately perform channel verification. Among them, the channel is the PCIE channel connecting the CPU (Central Processing Unit) and NTB in the controller.

In one implementation, the data to be backed up can be horizontally partitioned according to the data service object to obtain each zone; the data to be backed up can be vertically partitioned according to the importance of the data to obtain each vertical partition; based on the location of each zone in each vertical partition The amount of data allocated is a channel for each zone. For example, in an optional implementation, the backup data can be vertically partitioned according to the data importance to obtain vertical partitions of high importance, medium importance, and low importance. Each area can be divided into three vertical partitions. Channels are allocated to each area according to the amount of data. The larger the amount of data of high importance, the more channels are allocated. If the amount of data of high importance is the same, allocation will be based on the amount of data of medium importance. Obtain the data service object of each data in the data to be backed up from the front-end server; horizontally partition the data to be backed up according to the data service object to obtain each zone.

Among them, the data to be backed up is the current processing data of the controller. That is, the controller first partitions the current data processing. The partitioning method adopts a combination of horizontal partitioning and vertical partitioning. The horizontal partitioning divides the processing data horizontally according to data service objects. , the data service object is the object served by the data, which is provided by the front-end server. The controller allocates the processing data and data service objects. After the allocation is completed, the area where each data service object is located corresponds to its own processing data. The vertical partition divides the processing data vertically into three areas according to the importance: high, medium and low. Each area Includes a certain amount of processing data. After the partitioning is completed, PCIE channels are allocated to each area. The channels are allocated according to the previous horizontal partitioning situation. For example, if it includes 5 data service objects, it is divided into 5 areas. The number of channels allocated to each area is based on the importance of the data in the vertical partitioning. The data of the service object includes a large number of highly important data that require multiple allocations. PCIE channel.

Step S12: Continuously send data through all channels of each zone to the first non-transparent bridge connected to the main controller, so that the first non-transparent bridge can perform continuity check on the received data corresponding to each zone.

Step S13: Obtain the first verification result returned by the first non-transparent bridge. If the first verification result shows that a certain area fails the continuity check, perform enumeration verification on all channels in the area.

In one implementation, data is continuously sent to the first non-transparent bridge connected to the main controller through the channels in this area, so that the first non-transparent bridge performs continuity processing on the received data corresponding to each channel. check. That is, data is continuously sent to the first non-transparent bridge one by one through the channels in this area, so that the first non-transparent bridge can perform continuity verification on the received data to locate the problematic channel and cut off and isolate the problematic channel.

It can be seen that the embodiment of the present application first partitions the backup data and allocates channels to each zone, and then continuously sends data to the first non-transparent bridge connected to the main controller through all channels in each zone, so that the first non-transparent bridge Perform a continuity check on the received data corresponding to each zone, and then obtain the first check result returned by the first non-transparent bridge. If the first check result indicates that a certain zone has failed the continuity check, then Perform enumeration verification on all channels in this area. That is to say, before data backup, this application partitions and allocates channels for the backup data, and then verifies each area first. If a certain area fails the verification, then verifies each channel. In this way, it can accurately Verifying the problematic channels connected on the non-transparent bridge and handling the problems in a timely manner can improve the accuracy and efficiency of non-transparent bridge verification management, thereby ensuring the accuracy of data backup.

As shown in Figure 3, this embodiment of the present application discloses an optional non-transparent bridge management method, which is applied to the main controller in a dual controller, including:

Step S21: Partition the data to be backed up and allocate channels to each zone.

Step S22: Continuously send data to the first non-transparent bridge connected to the main controller through all channels of each zone, so that the first non-transparent bridge can perform continuity check on the received data corresponding to each zone.

Step S23: Obtain the first verification result returned by the first non-transparent bridge. If the first verification result shows that a certain area fails the continuity check, perform enumeration verification on all channels in the area.

Regarding the implementation of the above steps S21 to S23, reference may be made to the content disclosed in the foregoing embodiments, which will not be described again here.

Step S24: If the first verification result is that each zone passes the continuity check, data backup is started, and the data to be backed up is backed up to the backup controller in the dual controller according to each zone and through the channel of each zone. .

That is, after the channel allocation is completed, the controller sends the same continuous message data to NTB through the channel in each zone. NTB verifies the continuity of the message data in each zone in real time. If the data continuity in each zone is detected, After all requirements are met, the verification is completed and the data backup work begins. When an abnormality in data continuity is detected, NTB can quickly locate it. to a certain district. Then the channels in this area are enumerated and verified. After verifying the problematic channel, diagnostic information is sent to the controller. During the data backup process, the channel is cut off and isolated in software.

Moreover, in one implementation, the partition and channel allocation information can be sent to the backup controller, so that the backup controller can, based on the partition and channel allocation information, send data to the second server connected to the backup controller through all channels in each zone. The non-transparent bridge continuously sends data, so that the second non-transparent bridge performs a continuity check on the received data corresponding to each zone, and the second check result returned by the second non-transparent bridge is that a certain zone failed During continuity check, all channels in the area are enumerated and checked. Correspondingly, in one implementation, if the first verification result is that each area passes the continuity check, and the second verification result is that each area passes the continuity check, data backup is started.

That is, in order to ensure that the NTBs on both controllers on the backup link are available, the embodiment of this application also performs channel verification before backup on the backup controller side. The method can refer to the verification in the main controller. Moreover, in the embodiment of this application, the main controller and the backup controller in the dual controller are not fixed. The main controller is the controller where the data to be backed up is located during the backup process, and the backup controller is the controller to which the data to be backed up is backed up. controller.

In one implementation, the data to be backed up can be divided into batches to obtain multiple batches of data, and the multiple batches of data can be backed up to the backup controller in the dual controller one by one according to each zone and through the channel of each zone. ; Correspondingly, after the backup of any batch of data is completed, extract the first feature of the batch of data and transfer the first feature to the first non-transparent bridge, so that the first non-transparent bridge compares whether the first feature and the second feature are consistent. ; Wherein, the second feature is the second feature of the batch of data received by the backup controller extracted by the backup controller.

Optionally, in the embodiment of this application, extracting the first feature of the batch of data includes the following steps:

Step 00: Determine the hash function.

In one implementation, the data type of the batch of data can be identified; a hash function is determined based on the data type.

Among them, the data type includes text type, image type, and video type. Correspondingly, determining the hash function based on the data type includes: constructing the first hash function based on the division and leaving remainder method; determining the hash function of the text type as the first Hash function; construct a second hash function based on the square median method; determine the hash function of image type and video type as the second hash function.

Step 01: Use the hash function to extract the hash value set of the batch of data to obtain the first feature of the batch of data.

In one implementation, for any data type in the batch of data, a hash function corresponding to the data type can be used to extract a hash value; hash values corresponding to different data types are merged into the batch of data A set of hash values to obtain the first feature of the batch of data.

Furthermore, the main controller may send an execution instruction to the backup controller, so that the backup controller extracts the second feature of the received batch data under the control of the execution instruction. Under the control of executing instructions, the backup controller and the main controller perform feature extraction at the same time. The backup controller obtains the second feature, and the main controller obtains the first feature. Among them, the extraction of the second feature For the extraction process, reference can be made to the extraction process of the first feature, which will not be described again here. The main controller obtains the second characteristic sent by the backup controller; and transmits the second characteristic to the first non-transparent bridge, so that the first non-transparent bridge compares whether the first characteristic and the second characteristic are consistent.

That is to say, the embodiment of the present application establishes a hash value set for the processed data based on the Hash (hash) principle. The data amount of the hash value set is far smaller than the original processed data amount, and the original data and backup can be combined in the shortest time. Comparing data, the premise of establishing a hash value set is to select an appropriate hash function, so that the original data and the hash data mapped by the hash function have a many-to-one relationship, thus greatly reducing the amount of data processed. In the mapping process, the set of hash values mapped by the hash function needs to meet two conditions: 1. The total amount of data is much smaller than the total amount of original processed data, 2. The data in the hash set can reflect the entirety of the original data. In this case, the embodiment of the present application uses a hybrid hash function method, that is, different hash functions are selected for different data types, and finally the obtained sets are fused to obtain the final hash set. For example, for text type data, the division-leaving-remainder method is used, and the remainder after the text data segment is divided by the keyword P is used as the hash value, where P is a prime number smaller than the length of the data segment. This method reduces the It can greatly retain the characteristics of the original data while reducing the amount of data, and is suitable for data with relatively low complexity; for complex binary data such as images and videos, the square median method can be used to square the continuous data, and then select the middle of the resulting data. Several digits (i.e., the median and a preset number adjacent to the median) are used to extract a small number of feature values from the continuous data segment, thereby achieving dimensionality reduction of the data processed by the controller.

Optionally, if the first feature and the second feature are consistent, continue the backup of the next batch of data; if the first feature and the second feature are inconsistent, determine the problem channel, close the problem channel, and save the data of the problem channel. The relocation task is assigned to other channels in the area. Among them, the process of determining the problem channel includes: continuously sending data through all channels in each zone to the first non-transparent bridge connected to the main controller, so that the first non-transparent bridge receives the received data corresponding to each zone respectively. Perform continuity check; obtain the check result returned by the first non-transparent bridge. If the check result shows that a certain area fails the continuity check, all channels in the area will be enumerated and checked, and the failed The channel that passes the verification is determined to be the problem channel.

Moreover, the first feature can also be sent to the backup controller, and the second non-transparent bridge in the backup controller compares the first feature and the second feature to see whether they are consistent, and at the same time locates the problem of the corresponding channel of the second non-transparent bridge. That is, the two controllers simultaneously locate the problem channel.

Among them, the process of determining the problem channel in the backup controller is: continuously sending data through all channels in each zone to the second non-transparent bridge connected to the backup controller, so that the second non-transparent bridge can respond to each received The data corresponding to each area is checked for continuity. Obtain the verification result returned by the second non-transparent bridge. If the verification result shows that a certain area fails the continuity check, then enumerate and verify all channels in the area.

For example, see Figure 4, which is an optional data backup flow chart disclosed in an embodiment of the present application. The partition enumeration scanning algorithm implements data verification and channel management between the CPU (Central Processing Unit, Central Processing Unit) and the PCIE channel of the NTB. After the verification is correct, the data backup work starts. During the data backup process, the data is backed up based on The data comparison algorithm based on the Hash principle. The CPU side of the dual controller uses this algorithm to quickly perform feature extraction and verification on the backup data to ensure that the data is valid during the transmission process. When the above algorithm detects an error in the data backup during the data backup process, then Secondary boot partition enumeration scanning algorithm, this algorithm can quickly locate the abnormal PCIE channel and send an indication of the data transmission error to the BMC control unit as an exception log collection for the system, but the secondary boot does not perform partitioning and Divide channels and verify the channels that have been divided before backup. That is to say, the partition enumeration scanning algorithm performs verification processing on each PCIE channel at the NTB level. This algorithm is executed before data backup to determine the integrity of each channel. It is executed during the problem location analysis process for rapid Locate the problematic channel and implement NTB quick verification management operations through this algorithm. It can be understood that the partition scan enumeration scan algorithm can perform the following steps. Before data backup, partition the data to be backed up and allocate channels to each zone; pass the channel of each zone to the first non-stop connected to the main controller. The transparent bridge continuously sends data so that the first non-transparent bridge performs continuity verification on the received data corresponding to each zone; obtains the first verification result returned by the first non-transparent bridge. If the first verification result is a certain If the zone fails the continuity check, all channels in the zone will be enumerated and checked. For example, see FIG. 5 , which is a flow chart of an optional partition enumeration scanning algorithm disclosed in an embodiment of the present application. Moreover, the data comparison algorithm can be used to extract the first feature of any batch of data after the data backup is completed and transfer the first feature to the first non-transparent bridge, so that the first non-transparent bridge compares the first feature. and whether it is consistent with the second feature; wherein, the second feature is the second feature of the batch of data received by the backup controller extracted by the backup controller. In this way, the controller's CPU performs feature extraction on the real-time data transmitted during the data backup process, and then verifies the extracted feature data in a timely manner to ensure the accuracy of the data backup. Among them, after a data backup process, that is, a batch of data backup is completed, the main controller will issue an algorithm execution instruction to the backup controller. The dual controllers will execute the data comparison algorithm at the same time. In the end, both controllers will execute the algorithm to obtain The hash set is passed to the non-transparent bridge, and the non-transparent bridge verifies whether the two are completely consistent. For example, see FIG. 6 , which is a flow chart of an optional data comparison algorithm based on the Hash principle disclosed in the embodiment of the present application.

That is, the embodiment of the present application provides verification management before the normal operation of the NTB, as well as data verification diagnosis and exception handling solutions during the operation of the NTB. The purpose of the verification part before the normal operation of the NTB is to ensure that the NTB is operating normally from the main path to the converted path ends, so that the two controllers can back up data normally through the NTB channel. The channels on the NTB are partitioned and managed according to their functions through the partition enumeration scanning algorithm, and the enumeration algorithm is used to ensure that the channels in each zone are operating normally; after the NTB is operating normally, the characteristic part of the verification data is quickly extracted through the data comparison algorithm based on the Hash principle, and the accuracy of the mirror backup data is verified. The execution of the algorithm is accompanied by the completion of a batch of data backups in the data mirror backup between the dual controllers, ensuring that the backup continues after the verification is correct, which greatly avoids the data backup on the NTB. The error rate of backup is reduced. When data inconsistency is found during the execution of the above algorithm, the diagnosis process is started in time. The diagnosis process uses the partition enumeration scanning algorithm executed before NTB work as a benchmark to locate the NTB work area with problems, and then scans the problem channel through enumeration. The software promptly shuts down the data processing process on the channel, and transfers the data transmission task on the channel to other channels in the area. At the same time, an indication of the data transmission error is sent to the BMC (Baseboard Management Controller) control unit as the system's abnormal log collection. In this way, the NTB channel verification work before backup can be completed quickly, and when abnormal problems occur, appropriate responses can be made in time to minimize the impact of abnormal situations, thereby improving the efficiency of the storage system data backup channel and the safety and reliability of the system's work.

As shown in Figure 7, the embodiment of the present application discloses a non-transparent bridge management device, which is applied to the main controller in a dual controller, including:

The data partition module 11 is configured to partition the data to be backed up;

The channel allocation module 12 is configured to allocate channels to each zone;

The partition verification module 13 is configured to continuously send data through all channels of each zone to the first non-transparent bridge connected to the main controller, so that the first non-transparent bridge performs on the received data corresponding to each zone respectively. continuity check;

The verification result acquisition module 14 is configured to obtain the first verification result returned by the first non-transparent bridge;

The channel check module 15 is configured to perform an enumeration check on all channels in a certain area if the first check result is that a certain area fails the continuity check.

In one implementation, the data partition module 11 is configured to horizontally partition the data to be backed up according to the data service object to obtain each zone; to vertically partition the data to be backed up according to the importance of the data to obtain each vertical partition. Correspondingly, the channel allocation module 12 is configured to allocate channels to each region based on the data amount of each region in each vertical partition. Among them, the data partition module 11 is also configured to obtain data service objects from the front-end server. The channel distribution module 12 is configured to continuously transmit signals through the channel in this area to the first non-transparent bridge connected to the main controller. Send data so that the first non-transparent bridge performs continuity check on the received data corresponding to each channel.

Optionally, the device also includes a data backup module, which is configured to start data backup if the first verification result is that each zone passes the continuity check, and the data to be backed up will be processed according to each zone and passed through each zone. Channels are backed up to the backup controller in a dual controller.

Optionally, the device also includes a partition and channel allocation status sending module, which is configured to send the partition and channel allocation status to the backup controller after partitioning the data to be backed up and allocating channels to each zone, so that the backup controller can Based on the partition and channel allocation, data is continuously sent to the second non-transparent bridge connected to the backup controller through all channels in each zone, so that the second non-transparent bridge continuously performs continuity on the received data corresponding to each zone. Verify, and when the second verification result returned by the second non-transparent bridge is that a certain area fails the continuity check, enumerate and verify all channels in the area. Correspondingly, the data backup module is configured to start data backup if the first verification result is that each area passes the continuity check, and the second verification result is that each area passes the continuity check.

In one implementation, the data backup module is configured to divide the data to be backed up into batches to obtain multiple batches of data, and back up the multiple batches of data to dual controls one by one according to each zone and through the channel of each zone. backup controller in the server.

Correspondingly, the device also includes:

The first feature extraction module is configured to extract the first feature of any batch of data after the data backup is completed;

A first feature transmission module is configured to transmit the first feature to the first non-transparent bridge, so that the first non-transparent bridge compares whether the first feature and the second feature are consistent;

The second feature is the second feature of the batch of data received by the backup controller extracted by the backup controller.

Furthermore, the device further includes: an execution instruction sending module configured to send an execution instruction to the backup controller, so that the backup controller extracts the second characteristic of the received batch data under the control of the execution instruction.

Optionally, the device also includes a feature acquisition module configured to acquire the second feature sent by the backup controller;

The second feature transfer module is configured to transfer the second feature to the first non-transparent bridge, so that the first non-transparent bridge compares whether the first feature and the second feature are consistent.

Optional devices also include:

The problem channel determination module is configured to determine the problem channel if the first feature and the second feature are inconsistent;

The problem handling module is set to close the problem channel.

Among them, the problem channel determination module is configured to continuously send data to the first non-transparent bridge connected to the main controller through all channels in each zone, so that the first non-transparent bridge can respectively receive data corresponding to each zone. The data is checked for continuity; the check result returned by the first non-transparent bridge is obtained. If the check result is that a certain area fails the continuity If the verification is required, all channels in the area will be enumerated and verified, and the channels that fail the verification will be determined as problem channels.

In one implementation, the problem channel determination module includes the aforementioned partition verification module, verification result acquisition module, and channel verification module.

Optional, the first feature extraction module includes:

The hash function determination submodule is set to determine the hash function;

The first feature extraction sub-module is configured to use a hash function to extract a hash value set of the batch of data to obtain the first feature of the batch of data.

Among them, the hash function determines the sub-modules, including:

A data type identification unit is configured to identify the data type of the batch of data;

The hash function determination unit is configured to determine the hash function based on the data type.

The first feature extraction submodule includes:

The hash value extraction unit is configured to extract a hash value for any data type in the batch of data using a hash function corresponding to the data type;

The hash value fusion unit is configured to fuse hash values corresponding to different data types into a hash value set of the batch of data to obtain the first feature of the batch of data.

Optionally, the data type includes text type, image type, and video type. Correspondingly, the hash function determination unit is set to construct the first hash function based on the division-leaving-remainder method; determine the hash function of the text type as the third A hash function; constructing a second hash function based on the squared median method; determining the hash function of the image type and the video type as the second hash function.

Referring to Figure 8, the embodiment of the present application discloses a controller, including a processor 21 and a memory 22; the memory 22 is configured to save a computer program; the processor 21 is configured to execute the computer program to implement Following steps:

Partition the data to be backed up and allocate channels to each zone; continuously send data through all channels of each zone to the first non-transparent bridge connected to the main controller, so that the first non-transparent bridge receives each zone separately The corresponding data is subjected to continuity check; the first check result returned by the first non-transparent bridge is obtained. If the first check result is that a certain area fails the continuity check, all channels in the area are enumerated. check.

It can be seen that the embodiment of the present application first partitions the backup data and allocates channels to each zone, and then continuously sends data to the first non-transparent bridge connected to the main controller through all channels in each zone, so that the first non-transparent bridge Perform a continuity check on the received data corresponding to each zone, and then obtain the first check result returned by the first non-transparent bridge. If the first check result indicates that a certain zone has failed the continuity check, then Perform enumeration verification on all channels in this area. That is to say, before data backup, this application partitions and allocates channels for the data to be backed up, and then first Verify each area. If a certain area fails the verification, then verify each channel. In this way, the problematic channels connected to the non-transparent bridge can be accurately verified, and the problem can be dealt with in a timely manner, which can improve the non-transparent bridge verification. Management accuracy and efficiency, thereby ensuring the accuracy of data backup.

In this embodiment, when the processor 21 executes the computer subroutine stored in the memory 22, the following steps can be implemented: horizontally partition the backup data according to the data service object to obtain each area; vertically partition the backup data according to the importance of the data. , each vertical partition is obtained; channels are allocated to each zone based on the amount of data in each vertical partition.

In this embodiment, when the processor 21 executes the computer subroutine stored in the memory 22, the following steps can be implemented: If the first verification result is that each area passes the continuity check, the data backup is started, and the data to be backed up is Follow each zone and back up to the backup controller in the dual controller via each zone's channel.

In this embodiment, when the processor 21 executes the computer subroutine stored in the memory 22, the following steps can be implemented: after partitioning the data to be backed up and allocating channels to each zone, the partition and channel allocation information is sent to the backup controller. , so that the backup controller can continuously send data to the second non-transparent bridge connected to the backup controller through all channels of each zone based on the partition and channel allocation, so that the second non-transparent bridge can respond to the received data from each zone respectively. The corresponding data is subjected to continuity check, and when the second check result returned by the second non-transparent bridge is that a certain area fails the continuity check, all channels in the area are enumerated and checked.

In this embodiment, when the processor 21 executes the computer subroutine stored in the memory 22, the following steps can be implemented: If the first check result is that each area passes the continuity check, and the second check result is that each area passes the continuity check, If all zones pass the continuity check, data backup is started.

In this embodiment, when the processor 21 executes the computer subroutine stored in the memory 22, the following steps can be implemented: divide the data to be backed up into batches to obtain multiple batches of data, and merge the multiple batches of data one by one according to each area. Backup to the backup controller in the dual controller through the channel of each zone;

In this embodiment, when the processor 21 executes the computer subroutine stored in the memory 22, the following steps can be implemented: after the backup of any batch of data is completed, extract the first feature of the batch of data and transfer the first feature to the first non- The transparent bridge allows the first non-transparent bridge to compare whether the first feature and the second feature are consistent; wherein the second feature is the second feature of the batch of data received by the backup controller extracted by the backup controller.

In this embodiment, when the processor 21 executes the computer subroutine stored in the memory 22, the following steps can be implemented: sending an execution instruction to the backup controller, so that the backup controller extracts the first batch of data received under the control of the execution instruction. Two characteristics.

In this embodiment, when the processor 21 executes the computer subroutine stored in the memory 22, the following steps can be implemented: if the first feature and the second feature are inconsistent, determine the problem channel and close the problem channel.

In this embodiment, when the processor 21 executes the computer subroutine stored in the memory 22, the following steps can be implemented: determine the hash function; use the hash function to extract the hash value set of the batch of data to obtain the first hash value of the batch of data. feature.

In this embodiment, when the processor 21 executes the computer subroutine stored in the memory 22, the following steps can be implemented: identifying the data type of the batch of data; and determining a hash function based on the data type.

In this embodiment, when the processor 21 executes the computer subroutine stored in the memory 22, the following steps can be implemented: for any data type in the batch of data, use the hash function corresponding to the data type to extract a hash. value; fuse the hash values corresponding to different data types into a set of hash values of the batch of data to obtain the first feature of the batch of data.

In this embodiment, when the processor 21 executes the computer subroutine stored in the memory 22, the following steps can be implemented: constructing a first hash function based on the division-leaving-remainder method; determining a text-type hash function as the first hash function ; Construct a second hash function based on the squared median method; determine the hash function of the image type and the video type as the second hash function.

Moreover, the memory 22, as a carrier for resource storage, may be a read-only memory, a random access memory, a magnetic disk or an optical disk, etc., and the storage method may be short-term storage or permanent storage.

Optionally, the embodiment of the present application discloses a computer non-volatile readable storage medium, which is configured to save a computer program, wherein the following steps are implemented when the computer program is executed by the processor:

In this embodiment, when the computer subroutine stored in the computer non-volatile readable storage medium is executed by the processor, the following steps can be implemented: horizontally partition the data to be backed up according to the data service object to obtain each area; Vertically partition the backup data according to the importance of the data to obtain each vertical partition; allocate channels to each zone based on the amount of data in each vertical partition.

In this embodiment, when the computer subroutine stored in the computer non-volatile readable storage medium is executed by the processor, the following steps can be implemented: If the first check result is that each area passes the continuity check, start For data backup, the data to be backed up is backed up to the backup controller in the dual controller according to each zone and through the channel of each zone.

In this embodiment, when the computer subroutine stored in the computer's non-volatile readable storage medium is executed by the processor, the following steps can be implemented: after partitioning the data to be backed up and allocating channels to each zone, the partitions and channels are The allocation situation is sent to the backup controller, so that the backup controller can continuously send data to the second non-transparent bridge connected to the backup controller through all channels in each zone based on the partition and channel allocation situation, so that the second non-transparent bridge can respectively Perform a continuity check on the received data corresponding to each zone, and when the second check result returned by the second non-transparent bridge is that a certain zone fails the continuity check, enumerate all channels in the zone. Lift verification.

In this embodiment, when the computer subroutine stored in the computer non-volatile readable storage medium is executed by the processor, the following steps can be implemented: If the first check result is that each area passes the continuity check, and the If the second verification result is that each zone passes the continuity check, data backup is started.

In this embodiment, when the computer subroutine stored in the computer non-volatile readable storage medium is executed by the processor, the following steps can be implemented: divide the data to be backed up into batches, obtain multiple batches of data, and combine the multiple batches of data Back up to the backup controller in the dual controller batch by batch according to each zone and through the channel of each zone;

In this embodiment, when the computer subroutine stored in the computer non-volatile readable storage medium is executed by the processor, the following steps can be implemented: after the backup of any batch of data is completed, extract the first feature of the batch of data and store the first feature of the batch of data. A feature is transmitted to the first non-transparent bridge, so that the first non-transparent bridge compares whether the first feature and the second feature are consistent; wherein, the second feature is the first feature of the batch of data received by the backup controller extracted by the backup controller. Two characteristics.

In this embodiment, when the computer subprogram stored in the computer non-volatile readable storage medium is executed by the processor, the following steps can be implemented: sending an execution instruction to the backup controller so that the backup controller can extract it under the control of the execution instruction. Second characteristic of the received batch data.

In this embodiment, when the computer subroutine stored in the computer non-volatile readable storage medium is executed by the processor, the following steps can be implemented: if the first feature and the second feature are inconsistent, determine the problem channel, and assign the problem channel closure.

In this embodiment, when the computer subroutine stored in the computer non-volatile readable storage medium is executed by the processor, the following steps can be implemented: determine the hash function; use the hash function to extract the hash value set of the batch of data, Get the first feature of this batch of data.

In this embodiment, the computer subroutine stored in the computer non-volatile readable storage medium is executed by the processor. When , the following steps can be implemented: identify the data type of the batch of data; determine the hash function based on the data type.

In this embodiment, when the computer subroutine stored in the computer non-volatile readable storage medium is executed by the processor, the following steps can be implemented: for any data type of data in the batch of data, use the data corresponding to the data type The hash function extracts the hash value; the hash values corresponding to different data types are merged into the hash value set of the batch of data to obtain the first feature of the batch of data.

In this embodiment, when the computer subroutine stored in the computer non-volatile readable storage medium is executed by the processor, the following steps can be implemented: constructing a first hash function based on the division-leaving-remainder method; converting the text-type hash function Determine it as the first hash function; construct the second hash function based on the square median method; determine the hash function of the image type and the video type as the second hash function.

Each embodiment in this specification is described in a progressive manner. Each embodiment focuses on its differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other. As for the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple. For relevant details, please refer to the description in the method section.

The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be implemented directly in hardware, in software modules executed by a processor, or in a combination of both. Software modules may be located in random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or anywhere in the field of technology. Any other known form of non-volatile readable storage medium.

The above is a detailed introduction to a non-transparent bridge management method, device, controller and non-volatile readable storage medium provided by this application. This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and the core idea of the present application; at the same time, for those of ordinary skill in the art, there will be changes in the specific implementation and application scope based on the ideas of the present application. In summary, the contents of this specification should not be construed as limiting this application.

Claims

A non-transparent bridge management method, characterized in that it is applied to the main controller in a dual controller, including:

Partition the data to be backed up and allocate channels to each partition;

Data is continuously sent to the first non-transparent bridge connected to the main controller through all channels of each zone, so that the first non-transparent bridge performs continuity calibration on the received data corresponding to each zone. test;

Obtain the first verification result returned by the first non-transparent bridge. If the first verification result indicates that a certain area fails the continuity verification, then enumerate and verify all channels in the area.
The non-transparent bridge management method according to claim 1, wherein the step of partitioning the data to be backed up and allocating channels to each zone includes:

Horizontally partition the data to be backed up according to the data service object to obtain each zone;

Vertically partition the data to be backed up according to the importance of the data to obtain each vertical partition;

Channels are assigned to each zone based on the amount of data for each zone in each vertical partition.
The non-transparent bridge management method according to claim 2, wherein the data to be backed up is horizontally partitioned according to the data service object to obtain each area, including:

Get the data service object from the front-end server;

The data to be backed up is horizontally partitioned according to the data service object to obtain each area.
The non-transparent bridge management method according to claim 1, further comprising:

If the first verification result is that each zone passes the continuity check, data backup is started, and the data to be backed up is backed up to the dual server according to each zone and through the channel of each zone. Backup controller within controller.
The non-transparent bridge management method according to claim 4, characterized in that, after partitioning the data to be backed up and allocating channels to each partition, it further includes:

Send the partition and channel allocation information to the backup controller, so that the backup controller communicates through all channels in each zone to the second non-transparent controller connected to the backup controller based on the partition and channel allocation information. The bridge continuously sends data, so that the second non-transparent bridge performs continuity verification on the received data corresponding to each area, and the second verification result returned by the second non-transparent bridge is a certain area. When the continuity check fails, all channels in the area are enumerated and checked.
The non-transparent bridge management method according to claim 5, characterized in that if the first verification result is that each zone passes the continuity verification, then starting the data backup includes:

If the first verification result is that each zone passes the continuity verification, and the second verification result is that each zone passes the continuity verification, data backup is started.
The non-transparent bridge management method according to claim 1, characterized in that: all channels in the area Perform enumeration verification, including:

Channels in this area continuously send data to the first non-transparent bridge connected to the main controller, so that the first non-transparent bridge performs continuity check on the received data corresponding to each channel.
The non-transparent bridge management method according to claim 4, wherein the data to be backed up is backed up to the backup in the dual controller according to each zone and through the channel of each zone. Controllers, including:

The data to be backed up is divided into batches to obtain multiple batches of data, and the multiple batches of data are backed up to the backup in the dual controllers one by one according to each zone and through the channel of each zone. controller.
The non-transparent bridge management method according to claim 8, further comprising:

After the backup of any batch of data is completed, extract the first feature of the batch of data and transfer the first feature to the first non-transparent bridge, so that the first non-transparent bridge compares the first feature Whether it is consistent with the second characteristic;

Wherein, the second feature is a second feature of the batch of data received by the backup controller extracted by the backup controller.
The non-transparent bridge management method according to claim 9, further comprising:

An execution instruction is sent to the backup controller, so that the backup controller extracts the second feature of the received batch data under the control of the execution instruction.
The non-transparent bridge management method according to claim 10, further comprising:

Obtain the second characteristic sent by the backup controller;

The second characteristic is transferred to the first non-transparent bridge so that the first non-transparent bridge compares the first characteristic and the second characteristic for consistency.
The non-transparent bridge management method according to claim 9, further comprising:

If the first feature and the second feature are inconsistent, a problem channel is determined and the problem channel is closed.
The non-transparent bridge management method according to claim 12, characterized in that determining the problem channel includes:

Data is continuously sent to the first non-transparent bridge connected to the main controller through all channels of each zone, so that the first non-transparent bridge performs continuity on the received data corresponding to each zone. check;

Obtain the verification result returned by the first non-transparent bridge. If the verification result is that a certain area fails the continuity check, then enumerate and verify all channels in the area, and add the ones that have not passed the verification. The channel is determined to be the problem channel.
The non-transparent bridge management method according to claim 9, wherein the extracting the first feature of the batch of data includes:

Determine the hash function;

The hash function is used to extract the hash value set of the batch of data to obtain the first feature of the batch of data.
The non-transparent bridge management method according to claim 14, wherein the determining the hash function includes:

Identify the data type of the batch of data;

A hash function is determined based on the data type.
The non-transparent bridge management method according to claim 15, characterized in that, using the hash function to extract a hash value set of the batch of data to obtain the first feature of the batch of data includes:

For any data type in the batch of data, use the hash function corresponding to the data type to extract the hash value;

The hash values corresponding to different data types are merged into a set of hash values of the batch of data to obtain the first feature of the batch of data.
The non-transparent bridge management method according to claim 15, wherein the data type includes text type, image type, and video type. Correspondingly, determining the hash function based on the data type includes:

Construct a first hash function based on the division-leaving-remainder method;

Determine a text type hash function as the first hash function;

Construct a second hash function based on the squared median method;

A hash function of an image type and a video type is determined as the second hash function.
A non-transparent bridge management device, characterized in that it is applied to the main controller in a dual controller and includes:

The data partition module is set to partition the data to be backed up;

The channel allocation module is configured to allocate channels to each zone;

The partition verification module is configured to continuously send data to the first non-transparent bridge connected to the main controller through all channels of each zone, so that the first non-transparent bridge separately checks each received The data corresponding to the area is checked for continuity;

A verification result acquisition module configured to obtain the first verification result returned by the first non-transparent bridge;

The channel verification module is configured to perform enumeration verification on all channels in a certain area if the first verification result is that a certain area fails the continuity check.
A controller, characterized by including a memory and a processor, wherein:

The memory is configured to store a computer program;

The processor is configured to execute the computer program to implement the non-transparent bridge management method according to any one of claims 1 to 17.
A computer non-volatile readable storage medium, characterized in that it is configured to save a computer program, Wherein, when the computer program is executed by the processor, the non-transparent bridge management method according to any one of claims 1 to 17 is implemented.