TWI518502B - Apparatus, method and system for resilient computing cloud system - Google Patents

Description

Device, method and system for upgrading cloud computing cloud system

The present application relates to the field of storage computing technologies in the field of flexible computing cloud technology and block devices, and more particularly to an apparatus, method and system for upgrading an elastic computing cloud system that implements data storage based on a block device storage service.

The Elastic Computing Cloud is a cloud computing service that virtualizes local physical servers into multiple virtual servers based on virtualization technology to improve resource utilization and reduce costs. The storage service of the virtual server (ie, virtual machine) virtualized based on the flexible computing cloud technology is based on the remote storage service of the block device, which is called Elastic Block Storage in the industry.

The remote storage service of the block device is a block-level storage service provided by the Block Device Driver. The local physical server and/or virtual server can store the data of the server in a block-based manner to the remote storage resource pool, thereby saving local storage resources. In short, the remote storage service of the block device provides a service similar to the physical hard disk to store the data of the local physical server and/or virtual server, and has no format for the data format and file system. limit.

With the rapid development of flexible computing cloud technology, the elastic block storage service puts forward higher requirements for the high availability of remote storage services. Usually, virtual servers based on flexible computing cloud technology are tens of thousands of virtual servers, and the virtual servers that are virtualized are provided to third parties (such as government departments, traditional large and medium-sized enterprises). , small and medium-sized webmasters, individual users, etc.). The clusters of many virtual servers are all operated by the remote storage service of the block device. Obviously, the remote storage service carries and flows the data of the entire elastic computing cloud system, and the availability of the remote storage service is related to The availability of the entire elastic computing cloud system.

In the prior art, when an elastic computing cloud system that implements data storage based on a remote storage service of a block device is upgraded, it is necessary to first notify the service user that the service will be suspended due to the upgrade process, and then each virtual in the entire elastic computing cloud system The service of the server is suspended, so that the service upgrade of the storage service (such as the repair of the bug and the provision of new functions) can be performed. After the redeployed service works normally, the virtual machine in the entire elastic computing cloud system is served. The device resumes normal operation, completing an upgrade process. In addition, during the upgrade process, if the user is accessing the web page using the web service provided by the virtual server in the flexible computing cloud system, the user may not continue to access the web page, and if the user is using the virtual computing cloud system virtual system When the server performs decentralized operations, it also causes the user to terminate the decentralized operation currently in progress.

Therefore, in the prior art, when the flexible computing cloud system that implements data storage based on the remote storage service of the block device is upgraded, the upgrade process is very cumbersome, and the upgrade process takes a long time, and generally takes several hours. In order to complete the upgrade operation, the upgrade efficiency is relatively low, and the user has to interrupt the service currently being processed during the upgrade process. Therefore, it is not guaranteed to provide users with highly available services during the upgrade process, which reduces the user experience. .

The embodiments of the present invention provide an apparatus, a method, and a system for upgrading an elastic computing cloud system that implements data storage based on a block device storage service, thereby improving upgrade efficiency through online upgrade, and avoiding upgrading the service interruption used by the user. The resulting user experience is poor.

To solve the above problem, the embodiment of the present application provides a server in an elastic computing cloud system that implements data storage based on a block device storage service, including a block device driving device and at least one block device service device, where: a block device driving device, For receiving a data read/write request corresponding to a block device service device, placing the received data read/write request into a waiting queue pre-maintained for the block device service device, and extracting from the waiting queue The data read and write request is transferred to a pending queue pre-maintained for the block device service device, and the data read/write request in the pending queue is forwarded to the block device service device for processing, and the block is received After the device service device processes the received data read/write request, the forwarded data read/write request is deleted from the pending queue; and when it is determined that the device service device is to be upgraded, the pending The data read and write request in the queue is forwarded to the block device service for processing, and it is determined that the service upgrade or service will be performed. After the level is completed, the data read/write request in the pending queue is moved back to the waiting queue; the block device service device is configured to store the resource pool for the block device according to the data read/write request forwarded by the block device driving device. Perform read and write operations on the corresponding request data.

In order to solve the above problem, the embodiment of the present application further provides a method for upgrading an elastic computing cloud system that implements data storage based on a block device storage service, including each time a data read/write request corresponding to a block device service is received, Placing the received data read and write request into a waiting queue pre-maintained for the block device service; extracting the data read/write request from the waiting queue and migrating to the pending queue pre-maintained for the block device service And forwarding the data read/write request in the pending queue to the block device service for processing, and after receiving the processing result of the received device read/write request by the block device service, the data read and write request to be forwarded Deleting from the pending queue; and when it is determined that the block device service is to be upgraded, the data read and write request in the pending queue is forwarded to the block device service for processing; and the service is determined to be performed After the upgrade or business upgrade is completed, the data read and write requests in the pending queue are moved back to the waiting queue.

In order to solve the above problem, the embodiment of the present application further provides an apparatus for upgrading an elastic computing cloud system that implements data storage based on a block device storage service, including a waiting queue maintenance unit, configured to receive a corresponding block device each time. When the service data read and write request is made, the received data read/write request is placed in the waiting queue that is pre-maintained for the block device service; the pending maintenance unit is used to wait for maintenance from the waiting queue maintenance unit. The data read and write request is extracted from the column and is placed in a pending queue that is pre-maintained for the service of the block device; the read and write request processing unit is configured to read and write the data in the pending queue maintained by the pending queue maintenance unit. Forwarding to the block device service for processing, and after receiving the processing result of the received device read/write request by the block device service, deleting the forwarded data read/write request from the pending queue; and determining the When the block device service performs a service upgrade, the data read and write request in the pending queue is stopped from being forwarded to the block device service for processing, and Fix the service after the service upgrade or update is completed, and the read data queue pending requests column displaced back to the waiting queue.

To solve the above problem, the embodiment of the present application provides an elastic computing cloud system for implementing data storage based on a block device storage service, including a storage resource pool, configured to store data in an elastic computing cloud system based on a block device storage service manner; and at least An entity server is configured to: when receiving a data read/write request corresponding to a block device service, place the received data read/write request into a waiting queue that is pre-maintained for the block device service, and from the Waiting for the data read/write request to be extracted and relocated to the pending queue for the block device service, and reading the corresponding request data for the storage resource pool according to the data read/write request in the pending queue Write operation; and when it is determined that the block device service will perform a service upgrade, stop processing the data read and write request in the pending queue, and after the service upgrade or service upgrade is completed, the pending queue is The data read and write request is moved back to the wait queue.

The technical solution proposed by the embodiment of the present application maintains two queues for each virtual block device service in the entity server in the elastic computing cloud system that implements data storage based on the block device storage service, and one is data. The Waiting and Waiting Waiting queue is used to maintain data read and write requests for a service to a block device and wait for its processing. The other is a Pending queue for maintaining the physical server pin pair. The data read/write request currently being processed by the device service stops processing the data read/write request in the Pending queue maintained for the block device service when the device service needs to be upgraded, and the service upgrade is completed in the block device service. After the operation resumes, the data read and write request being processed in the Pending queue is again put into the Waiting queue to re-accept the scheduling process, thereby realizing the completion of the storage service in the elastic computing cloud system without interrupting the storage service. Service upgrade processing, which improves the efficiency of service upgrades and avoids users being used due to upgrades. Interruption, to provide users with high availability service during the upgrade process, thereby improving the user experience.

In the prior art, when the elastic computing cloud system that implements data storage based on the block device storage service is upgraded, since there is no hot deployment to implement the online upgrade mode, the virtual server in the elastic computing cloud system needs to be shut down. After the service upgrade, the virtual server is started after the upgrade is completed, which greatly increases the maintenance cost for the storage service provider and the storage service user, and greatly reduces the service availability. To solve this problem, the present application proposes a hot deployment of a block device storage service in an elastic computing cloud system that implements data storage based on a block device storage service, so as to implement an online upgrade storage service, that is, a storage service without interrupting the storage service. The upgrade process is to improve the availability of the service provider storage service and enhance the transparency of the storage service, so that the user of the storage service does not pay attention to the upgrade process of the backend service.

The hot deployment of the block device storage service is the online upgrade of the block device service. Since the block device storage service is based on the block device driving technology, the physical server in the elastic computing cloud system includes the kernel state. The block device driver service program, the kernel mode block device driver service program needs to forward the data read and write request to the user mode block device service program included in the entity server, and the user mode block device service program forwards the read and write request to the far The network storage resource pool or the local physical disk storage resource pool and so on are processed. Among them, the kernel mode block device driver service program is more stable than the user mode block device service program; and the user mode block device service program requires frequent version upgrades to provide better storage services. According to the technical solution provided by the embodiment of the present application, a block device driver service program that can support online upgrade is provided in a physical server in the flexible computing cloud system to improve the availability of the storage service and improve the user experience.

As shown in FIG. 1 , a topological structure diagram of an elastic computing cloud system based on a block device storage service implemented by a technical solution proposed by an embodiment of the present application, wherein the elastic computing cloud system includes several elastic computing cloud technologies Virtual server virtualized by the local physical server (virtual server 1, virtual server 2, virtual server 3.. virtual server n), each virtual server will The locally provided data storage service based on the block device storage service technology is implemented by using a remote storage resource pool or a local storage resource pool, that is, each virtual server stores the data that needs to be saved locally to the remote storage resource pool. Or the locally set storage resource pool, and when there is a request for reading data, the related data is read from the remote storage resource pool or the locally set storage resource pool. In this way, each virtual server can share a remote or locally set storage resource pool without occupying the storage resources of the local entity server itself, thereby improving the resource utilization of the elastic computing cloud system.

In order to implement the technical solution proposed by the present application, the function of the physical server in the flexible computing cloud system introduced above needs to be improved to achieve the purpose of online upgrade of the storage service.

As shown in FIG. 2, a schematic structural diagram of an improved physical server proposed in the embodiment of the present application includes a kernel device-based block device driving device 20, a user-mode user device 21, and at least one block device service device 22. And a local or remote storage resource pool 23, wherein the block device driver 20 can control the operation of the plurality of block device service devices 22, and each component implements an online upgrade of the storage service through the following working principle: a kernel device block device The driving device 20 receives the data read/write request from the read/write scheduling layer at the bottom of the physical server, and forwards the data read/write request to the block device service device 22 in the user state through, for example, the TCP protocol. In order to implement the online upgrade function supporting hot deployment, the block device driver 20 maintains two queues for each block device service device 22 virtualized by the physical server, one of which is a Waiting queue, and the block device driver Each time the device 20 receives the data read/write request corresponding to a certain block device service device 22 sent by the read/write scheduling layer of the underlying physical server, the device 20 puts the received read/write request into the device for the block device. The device 22 waits for the subsequent processing in the Waiting queue maintained by the device 22; the other is the Pending queue, and the block device driving device 20 firstly processes the data read and write request of the block device service device 22 from the block. The device service device 22 extracts the data read/write request from the previously maintained Waiting queue and moves it to the Pending queue pre-maintained for the block device service device 22, and forwards the request in the Pending queue to the block device service. The device 22 performs processing, and it can be seen that the read and write requests in the pending queue are the read and write requests currently being processed by the physical server, and the block device service device 22 is correctly processed. After the read/write request of the pending queue forwarded by the device driver 20, the result message of the correct processing is sent to the block device driving device 20. After the block device driving device 20 receives the result message of the correct processing, the processing will be forwarded. A good read and write request is removed from the unresolved queue, and then the other read and write requests in the pending queue are processed. Subsequently, when the block device driving device 20 determines that the block device service device 22 is to perform the service upgrade, the data read/write request in the Pending queue pre-maintained for the block device service device 22 is stopped from being forwarded to the block device service device 22. After processing, and after further determining that the service upgrade or service upgrade is to be completed, the block device driver 20 re-joins the request in the Pending queue pre-maintained by the block device service device 22 to serve the block device. 22 The pre-maintained Waiting queue waits for subsequent re-calls and processing.

Preferably, the block device driving device 20 can determine whether the block device service device 22 will perform an upgrade operation or has completed the upgrade by detecting whether the TCP communication connection between the user and the user mode block device service device 22 is normal. After detecting that the TCP communication connection between itself and the block device service device 22 is disconnected, it can be determined that the block device service device 22 will perform the service upgrade, and after detecting the TCP communication connection between itself and the block device service device 22, the connection is restored. It can be determined that the service upgrade of the block device service device 22 has been completed.

The user-mode block device service device 22 receives the data read/write request forwarded by the block device driving device 20 from the kernel state through, for example, the TCP protocol, that is, the corresponding block device service device 22 processed by the receiving block device driving device 20 is pre- The data read and write request in the maintained Pending queue, after receiving the read and write request, parses the request, and the parsing includes verifying whether the read/write request is legal, specifically a read request or a write request, and the read and write request Start position and length information, etc., and perform data read or write operations to the local or remote storage resource pool according to the analysis result. If it is a read request, the requested data is read from the local or remote storage resource pool 23, and the read data and the processing result are returned to the kernel device mode device driving device 20 through, for example, the TCP protocol. If it is a write request, the data to be written is stored in the local or remote storage resource pool 23, and the processing result of the write request is returned to the kernel device mode device driver 20 via, for example, the TCP protocol.

The user-side client device 21 can control the respective block device service device 22 to start or stop working through the kernel device driver device 20. When the control block device service device 22 starts the work, after the user initiates the request to create the block device service through the client device 21, the client device 21 first establishes a TCP connection with the block device service device 22, and then creates the created TCP Socket. The control code is passed to the block device driver 20 through the ioctl command, after which the client device 21 can close the TCP Socket socket, so that the client device 21 implements the block device driver 20 and the user mode block in the kernel mode. A TCP connection is established between the device service devices 22, and all subsequent data read and write requests to the block device drive device 20 can be forwarded to the block device service device 22 via the TCP connection. When the control block device service device 22 stops working, after the user initiates a request to stop a certain block device service through the client device 21, the client device 21 initiates a stop request to the block device driving device 20 through the ioctl command, and the block device driving device After receiving the request, the device first sends a stop request to the block device service device 22, and the block device service device 22 closes the TCP Socket socket after receiving the request and exits safely; the block device driver 20 receives the block device service device. After the stop reaction of 22, the read and write requests in the pending column are stopped, so that the block device service device 22 stops working. When the block device service device 22 is in the process of upgrading, the client device 21 may attempt to establish a TCP link with the block device service device 22 every second. If the new TCP link is successfully established, it indicates that the block device service device 22 is upgraded, and then the client device 21 transmits the newly created TCP Socket control code to the block device driving device 20 through the ioctl command. After the upgrade of the block device service device 22 is completed, the block device drive device 20 establishes a new TCP link with the block device service device 22 to perform communication processing of the data read/write request.

Of course, the client device 21 is an optional component in the physical server of the application. In addition to using the client device 21 to implement the user to control the working state of the user device device 22, it may also be based on other devices. The triggering mechanism such as the timing function and the specific event triggering function automatically controls the working state of the user equipment block device service device 22.

The local or remote storage resource pool 23, in general, the storage resource pool 23 is a storage server similar to the distributed file system, and the resource pool 23 can be deployed locally or remotely, in most cases A storage resource pool deployed remotely. The storage resource pool 23 stores service data in each virtual server in the elastic computing cloud system. Specifically, when receiving the data read request of a certain block device service device 22 in the physical server, the related request data is read from itself, and the read data is fed back to the block device service device 22 When a material write request of a block device service device 22 in the physical server is received, the data requested to be written is stored in itself.

As shown in FIG. 3, a schematic diagram of an internal processing flow for reading and writing data from a local or remote storage resource pool in the improved physical server in the embodiment of the present application, wherein the specific process of reading data is: when the local entity server The file system layer in the file system layer sends the data read request corresponding to a block device service device to the block device driver through the common block layer and the W/R scheduling layer, and the block device driver device puts the request into the block device service. Waiting for subsequent processing in the Waiting queue maintained by the device, and subsequently extracting the data read request from the Waiting queue into the Pending queue pre-maintained for the block device service device and processing the read request Forwarding the read request to the block device service device, the block device service device reads the related request data from the storage pool according to the read request, and returns the read result to the block device driving device through the TCP connection, the block device The driving device further forwards the reading result to the file system layer through the W/R scheduling layer and the common block layer, and the file system layer returns the reading result to the system call. .

The specific process of writing data is: the local entity server stores the data write request and the requested data through the file system layer, the common block layer, and the W/R scheduling layer in the system call to the block device driving device, and the block device. The driving device puts the write request into the Waiting queue pre-maintained by the block device service device, and waits for subsequent processing, and subsequently extracts the data write request from the Waiting queue into the service device for the block device. The pre-maintained Pending queue processes the write request, that is, forwards the data write request and the data to be written to the block device service device, and the block device service device implements the write request by parsing the data write request. The data is saved to the storage resource pool.

As shown in FIG. 4, it is a flowchart of a specific working principle of a block device driving device in an improved physical server according to an embodiment of the present application, wherein the block device driving device separately pre-maintains two for each block device service device. The queue is a Waiting queue, which is used to place a data read and write request sent to the corresponding block device service device and wait for its processing, so that it enters the waiting queue, and the other queue For placing data read and write requests currently being processed for the corresponding block device service device, based on the two queues of maintenance, the specific working principle of the block device driving device is as follows: Step 40: The block device driving device receives each time When the data read/write request for a block device service device is sent to the bottom of the physical server, the received data read/write request is placed in the Waiting queue pre-maintained for the block device service device; step 42 When the block device driver processes the data read/write request corresponding to the block device service device, first extracts the data read/write request from the Waiting queue and moves Going to the Pending queue pre-maintained for the block device service device, and forwarding the data read/write request in the Pending queue to the device service device in the user mode for processing, and receiving the block device service device pair After the received data read/write request is correctly processed, the forwarded data read/write request is deleted from the Pending queue; in step 44, the block device driver stops when the device service device is to be upgraded. All data read and write requests in the Pending queue are forwarded to the block device service device for processing; in step 46, after the block device driver determines that the service upgrade or service upgrade is completed, the data in the Pending queue is read and written. The request is then relocated back to the Waiting queue to continue waiting for subsequent calls to be called and processed. Specifically, the block device driving device may determine whether the block device service device will perform a service upgrade or an upgrade end by detecting a communication connection state between itself and the block device service device. More specifically, after detecting that the communication connection between the user and the block device service device is disconnected, it may be determined that the block device service device will perform a service upgrade, and detects a communication connection between itself and the block device service device. After the connection is restored, it can be determined that the block device service device has completed the service upgrade.

As shown in FIG. 5, the block device driving device in the physical server improved in the embodiment of the present invention performs a process of scheduling the maintenance of the Waiting queue and the Pending queue before and after the service upgrade of the block device service device. In order to support the hot deployment (ie, online upgrade) of the block device service device, the kernel device device driver needs to maintain the status of the received data read/write request. In the block device driving device, two data read/write queues are respectively used for each block device service device to maintain the state of the read/write request, and one is a Waiting Queue for serving each block device. Maintain W/R requests sent by the W/R scheduling layer in the physical server. These requests are queued in the Waiting queue for subsequent processing by the block device driver; the other is the pending Pending Queue. Maintaining the block device driver for each W/R request that the block device service device is currently processing, the W/R requests in the pending queue are not processed, and may be in the block device driver waiting for the processing of the corresponding block device service device. Response status. When a block device service device does not perform a service upgrade, the block device driver first takes a W/R from the head of the Waiting queue that is pre-maintained for the block device service device when processing the W/R request. Requesting, and deleting the fetched W/R request from the Waiting queue, and then adding the W/R request to the head of the Pending queue pre-maintained for the block device service device, and the Pending伫Each W/R request in the column is forwarded to the block device service device for processing, and after receiving the result that the block device service device correctly processes the received data read/write request, the forwarded data read/write request is sent from Pending伫Deleted in the column. When the device service device of the device needs to be upgraded, the block device driver stops forwarding the W/R request in the corresponding Pending queue to the block device service device for processing, that is, no new device is sent to the block device service device. W/R request. After the block device service device is upgraded or upgraded, the block device driver will respond to all W/R requests in the corresponding Pending queue (as shown in Figure 5, requests 1, 2, 3, 4, 5, 6....) Rejoin the corresponding Waiting queue to re-receive the rescheduling process of the block device driver.

It can be seen that by adding a block device driving device to the physical server in the elastic computing cloud system, and scheduling the two queues pre-maintained for each block device service device through the block device driving device, The storage service of the background storage service is upgraded while the storage service is provided for the data storage user, thereby improving the availability of data storage in the flexible computing cloud system. In addition, the above solution is transparent to the user, so that the data storage user can hardly perceive the impact of the upgrade of the storage service service on the data storage service, thereby improving the user experience.

Correspondingly, as shown in FIG. 6, a specific structural diagram of a block device driving device added to the physical server improved in the embodiment of the present application, specifically includes a waiting queue maintenance unit 60, which is used for each block respectively. The device service device pre-maintains a Waiting queue for placing each received data read/write request for the corresponding block device service device to enter a waiting queue state; the pending queue maintenance unit 61 is configured to Each block device service device pre-maintains another Pending queue for placing a data read/write request currently being processed for the corresponding block device service device; specifically, a service device for a block device from the Waiting queue maintenance unit 60 Extracting the data read/write request from the maintained Waiting queue and migrating it to the Pending queue pre-maintained for the block device service device; specifically, the pending queue maintenance unit 61 may be from the corresponding Waiting queue Extracting the data read/write request, and deleting the extracted data read/write request in the Waiting queue, and then shifting the extracted data read/write request Go to the header of the Pending queue that corresponds to the pre-maintenance. The read/write request processing unit 62 is configured to forward the data read/write request in the corresponding pending queue maintained by the Pending queue maintenance unit 61 to the block device service device for processing, and receive the received by the block device service device. After the data read and write request is correctly processed, the forwarded data read and write request is deleted from the Pending queue; and when it is determined that the block device service device in the elastic computing cloud system will perform the service upgrade, the Pending queue is stopped. The data read/write request is forwarded to the block device service device for processing, and after determining that the service upgrade or the service upgrade is completed, the data read/write request in the corresponding Pending queue is moved back to the corresponding Waiting queue to continue. It is awaiting subsequent scheduling processing by the read/write request processing unit 62. Specifically, the read/write request processing unit 62 may specifically detect a communication connection state between the block device driving device and the block device service device, and after detecting that the communication connection is disconnected, determine that the block device service device performs a service upgrade, and After detecting that the communication connection is restored, it can be determined that the service upgrade of the block device service device is completed.

As shown in FIG. 7 , in the embodiment of the present application, when the server service is upgraded, the operation of the block device driving device in the physical server is controlled. The operation process is as follows: Step 70: Before the storage service upgrade, upgrade The maintenance personnel issues an ioctl command to the kernel mode block device driver through the user mode client device, instructing the block device driver to hold all W/R requests.

In step 72, the upgrade maintenance personnel perform service upgrades on the storage service according to the normal upgrade process (service upgrade includes stopping the service, upgrading the service program, releasing a new service, etc.). This completes a hot deployment of online service upgrades to the storage service, and during the hot deployment process, the block device driver will hold all W/R requests.

In step 74, when deployed, the block device driver releases all W/R requests and the data storage service will continue to be processed.

Those skilled in the art will appreciate that embodiments of the present application can be provided as a method, apparatus (device), or computer program product. Thus, the present application can take the form of a fully hardware embodiment, a fully software embodiment, or an embodiment combining the software and hardware. Moreover, the present application can take the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer usable code therein. .

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus, and computer program products according to embodiments of the present application. It will be understood that each flow and/or block of the flowchart illustrations and/or FIG. These computer program instructions can be provided to a processor of a general purpose computer, a special purpose computer, an embedded processor or other programmable data processing device to produce a machine for generating instructions for execution by a processor of a computer or other programmable data processing device. Means for implementing the functions specified in one or more of the flow or in a block or blocks of the flow chart.

The computer program instructions can also be stored in a computer readable memory that can boot a computer or other programmable data processing device to operate in a particular manner, such that instructions stored in the computer readable memory produce an article of manufacture including the instruction device. The instruction means implements the functions specified in one or more blocks of the flow or in a flow or block diagram of the flowchart.

These computer program instructions can also be loaded onto a computer or other programmable data processing device to perform a series of operational steps on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device. The instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.

Although the preferred embodiment of the present application has been described, those skilled in the art can make additional changes and modifications to the embodiments once they are aware of the basic inventive concept. Therefore, the scope of the appended claims is intended to be construed as a

It will be apparent to those skilled in the art that various modifications and changes can be made in the present application without departing from the spirit and scope of the application. Thus, it is intended that the present invention cover the modifications and variations of the present invention.

20. . . Block device driver

twenty one. . . Client device

twenty two. . . Block device service device

twenty three. . . Local or remote storage resource pool

60. . . Waiting for the maintenance unit

61. . . Pending maintenance unit

62. . . Read and write request processing unit

FIG. 1 is a topological structural diagram of an elastic computing cloud system for implementing data storage based on a block device storage service according to a technical solution provided by an embodiment of the present application;

2 is a schematic structural diagram of a structure of an improved physical server according to an embodiment of the present application;

3 is a schematic diagram of an internal processing flow of an improved physical server reading and writing data to a local or remote storage resource pool in the embodiment of the present application;

4 is a flowchart of a specific working principle of a block device driving device in an improved physical server according to an embodiment of the present application;

FIG. 5 is a schematic diagram of processing of scheduling a maintained Waiting queue and a Pending queue before and after a service upgrade of a block device service device according to an embodiment of the present invention;

6 is a schematic structural diagram of a specific block device driving device in an improved physical server according to an embodiment of the present application;

FIG. 7 is a schematic diagram of the operation of the block device driving device in the control entity server when the server service is upgraded in the embodiment of the present application.

Claims (10)

A server in an elastic computing cloud system for implementing data storage based on a block device storage service, comprising: a block device driving device and at least one block device service device, wherein: the block device driving device is configured to receive a corresponding one When the block device service device reads and writes a data request, the received data read/write request is placed in a waiting queue pre-maintained for the block device service device, and the data read/write request is extracted and displaced from the waiting queue. Going to the pending queue pre-maintained for the block device service device, and forwarding the data read/write request in the pending queue to the block device service device for processing, and receiving the block device service device to receive the received After the processing result of the data read/write request, the forwarded data read/write request is deleted from the pending queue; and when it is determined that the device service device is to be upgraded, the reading and reading of the data in the pending queue is stopped. Request to forward to the block device service for processing, and after determining that the service upgrade or service upgrade is completed, the pending The data read and write requests displaced back to the waiting queue; block device service apparatus, for forwarding to the data read or write request block device driver apparatus according to the corresponding request data read and write operations for a block storage device resource pool. The server of claim 1, further comprising a client device for controlling the startup of the block device service device, and establishing between the device driver device and the activated device device after startup The communication connection; and the control block device drive stop, and after the stop, the communication connection between the block device drive and the block device service device is removed. The server of claim 1, wherein the block device driving device specifically detects a communication connection between itself and the block device service device, and detects between itself and the block device service device. After the communication connection is disconnected, it is determined that the device service device of the device is to be upgraded, and after detecting the communication connection between the device and the device device, the service upgrade of the device service device is completed. A method for upgrading an elastic computing cloud system for implementing data storage based on a block device storage service, comprising: reading and receiving received data each time a data read/write request corresponding to a block device service is received Requesting to be placed in a wait queue pre-maintained for the block device service; extracting data read and write requests from the wait queue and migrating to a pending queue pre-maintained for the block device service, and placing the pending queue The data read/write request is forwarded to the block device service for processing, and after receiving the processing result of the received device read/write request by the block device service, the forwarded data read/write request is deleted from the pending queue And when it is determined that the device service is to be upgraded, the data read and write request in the pending queue is forwarded to the block device service for processing; and after it is determined that the service upgrade or service upgrade is completed, The data read and write request in the pending queue is moved back to the wait queue. The method of claim 4, wherein the device service is determined to be upgraded and the service upgrade is completed, and the method includes: detecting a communication connection with the device service; After the communication connection between the block device services is disconnected, it is determined that the block device service will perform a service upgrade; and after detecting the communication connection with the block device service to restore the connection, it is determined that the block device service service upgrade is completed. The method of claim 4, wherein the data read/write request is extracted from the waiting queue and moved to a pre-maintained pending queue, specifically: from the queue header of the waiting queue Extracting the data read/write request; deleting the extracted data read/write request in the waiting queue; and shifting the extracted data read/write request into the pre-maintained pending queue header. An apparatus for upgrading an elastic computing cloud system that implements data storage based on a block device storage service, comprising: waiting for a queue maintenance unit, configured to receive a data read/write request corresponding to a block device service each time And the received data read/write request is placed in a waiting queue that is pre-maintained for the block device service; the pending queue maintenance unit is configured to extract the data read/write request from the waiting queue maintained by the waiting queue maintenance unit. And being moved to a pending queue that is pre-maintained for the service of the block device; the read/write request processing unit is configured to forward the data read/write request in the pending queue maintained by the pending queue maintenance unit to the block device service. Processing, and after receiving the processing result of the received device read/write request by the block device service, deleting the forwarded data read/write request from the pending queue; and determining that the block device service will perform service upgrade Stopping the data read and write request in the pending queue to the block device service for processing, and determining that the service upgrade will be performed or After the service upgrade is complete, the column read and write data queue pending request displaced back to the waiting queue. The device of claim 7, wherein the read/write request processing unit specifically determines a communication connection between the device and the block device service, and after detecting that the communication connection is disconnected, determining the block. The device service will perform a service upgrade, and after detecting that the communication connection is restored, it is determined that the service upgrade of the device service is completed. The device of claim 7, wherein the pending queue maintenance unit specifically extracts a data read/write request from the queue header of the waiting queue, and reads the extracted data in the waiting queue The write request is deleted, and the extracted data read and write request is moved to the header of the pre-maintained pending queue. An elastic computing cloud system for implementing data storage based on a block device storage service, comprising: a storage resource pool, configured to store data in an elastic computing cloud system based on a block device storage service manner; and at least one physical server, Each time a data read/write request corresponding to a block device service is received, the received data read/write request is placed in a waiting queue pre-maintained for the block device service, and data is extracted from the waiting queue. Read and write the request and move it to the pending queue that is pre-maintained for the block device service, and according to the data read and write request in the pending queue, perform read and write operations on the corresponding request data for the storage resource pool; When the device service is upgraded, the data read/write request in the pending queue is stopped, and the data read/write request in the pending queue is moved after the service upgrade or the service upgrade is completed. Set back to the wait queue.