CN109814946B - A method and device for realizing distributed storage in XenServer - Google Patents

Abstract

The present invention provides a method and device for realizing distributed storage in XenServer. The method includes: receiving a user-triggered storage repository establishment message, and using the storage pool established in the distributed storage cluster as the storage repository of XenServer; receiving The physical block device PBD creation message triggered by the user determines all XenServer host information carried in the PBD creation message, creates a PBD for connecting each XenServer host and the repository, and mounts all the created PBDs; Disk image VDI creation message, VDI is created in XenServer and stored in the storage repository; virtual disk mapping message triggered by the user is received, and the mapping relationship between each block device in the distributed storage cluster and each VDI in the storage repository is established. The present invention enables XenServer to support distributed storage.

Description

A method and device for realizing distributed storage in XenServer

technical field

The present invention relates to the technical field of storage, in particular to a method and device for realizing distributed storage in XenServer.

Background technique

XenServer is a server virtualization platform launched by Citrix. The software package contains Xen open source, paravirtualized hypervisor and nearly perfect Xen open source running on virtual x86 computers that need to create and manage deployments. performance. XenServer is specially optimized for Windows and Linux. It does not need the underlying operating system and can be directly run on server hardware. It is a system with high efficiency and high scalability.

In the existing implementation, XenServer does not have distributed storage, which is the most fatal for Citrix XenServer in hyper-converged HCI, and is not conducive to the creation of a "hyper-converged" all-in-one machine based on XenServer.

SUMMARY OF THE INVENTION

In view of this, the purpose of the present invention is to provide a method and apparatus for realizing distributed storage in XenServer, so that XenServer can support distributed storage.

In order to achieve the above object, the present invention provides the following technical solutions:

A method for implementing distributed storage in XenServer, the method includes:

Receive the repository creation message triggered by the user, and use the storage pool created in the distributed storage cluster as the repository of XenServer;

Receive the physical block device PBD creation message triggered by the user, determine all XenServer host information carried in the PBD creation message, create a PBD for connecting each XenServer host and the storage library, and mount all the created PBDs;

Receive a user-triggered virtual disk image VDI creation message, create a VDI in XenServer and store it in the repository;

A virtual disk mapping message triggered by a user is received, and a mapping relationship between each block device in the distributed storage cluster and each VDI in the storage library is established.

A plug-in for realizing distributed storage in XenServer, the plug-in includes: a receiving unit and a processing unit;

The receiving unit is configured to receive a user-triggered repository establishment message; a user-triggered physical block device PBD creation message; a user-triggered virtual disk image VDI creation message; a user-triggered virtual disk creation message map message;

The processing unit is used for the receiving unit to use the storage pool established in the distributed storage cluster as the storage library of XenServer when the receiving unit receives the user-triggered storage repository establishment message; for the receiving unit to receive the user-triggered PBD creation message When, determine all the XenServer host information carried by the PBD creation message, create a PBD for connecting each XenServer host and the repository, and mount all the created PBDs; when the receiving unit receives the user-triggered VDI creation message , creating a VDI in the XenServer and storing it in the repository; when the receiving unit receives a user-triggered virtual disk mapping message, establishes a mapping relationship between each block device in the distributed storage cluster and each VDI in the repository.

As can be seen from the above technical solutions, in the present invention, a storage pool is established in a distributed storage cluster, and the storage pool is used as a storage library of XenServer; the storage library and the host in XenServer are connected by creating and mounting PBDs. , and then create multiple VDIs in the repository, and establish the mapping relationship between each VDI in the repository and each block device in the distributed storage cluster, so as to introduce the distributed storage cluster into XenServer and implement XenServer Support for distributed storage.

Description of drawings

1 is a flowchart of a method for implementing distributed storage in XenServer according to an embodiment of the present invention;

FIG. 2 is a schematic structural diagram of a plug-in for implementing distributed storage in XenServer according to an embodiment of the present invention.

Detailed ways

In order to make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be described in detail below with reference to the accompanying drawings and embodiments.

Referring to FIG. 1, FIG. 1 is a flowchart of a method for implementing distributed storage in XenServer according to an embodiment of the present invention. As shown in FIG. 1, the method includes the following steps:

Step 101: Receive a user-triggered repository establishment message, and use the storage pool established in the distributed storage cluster as the XenServer repository.

In this embodiment of the present invention, the distributed storage cluster is a Ceph-based distributed storage cluster, and a storage pool for storing VDI (Virtual Disk Image) may be created in the distributed storage cluster in advance. After the plug-in is installed on XenServer, the user can trigger a repository creation message in XenServer. After receiving the repository creation message in the background, the storage pool can be used as the XenServer repository.

In practical applications, an application program interface for creating a repository can be pre-configured. The input parameters of the application program interface are storage pool information (such as the uuid of the storage pool), attribute information of the repository (such as the blocks supported by the repository) device type), the main function is to create a repository based on the storage pool specified by the input parameter according to the sr-introduce command provided by XenServer, and have the repository properties specified by the input parameter.

The following is an example of executing the repository build command (sr-introduce) provided by XenServer:

#xe sr-introduce name-label="CEPH RBD Storage" type=rbd uuid=cda789de-c8d3-4d0d-8510-7deaa8c53aa8shared=true content-type=user

In the above command, cda789de-c8d3-4d0d-8510-7deaa8c53aa8 is the uuid of the storage pool. After the above command is executed, a repository based on the storage pool is created, and the uuid of the repository is obtained (assuming that the repository uuid is 4ceb0f8a-1539-40a4-bee2-450a025b04e1).

In the above command, the block device type supported by the repository (one of the repository properties) is set to rbd (ie type=rbd), indicating that the block device type corresponding to the VDI stored in the repository is rbd.

Therefore, in this step, the user can specify the storage pool and the storage library attributes in the trigger interface of the storage library establishment message provided by XenServer, and then trigger the storage library establishment message (carrying the specified storage pool information and storage library attribute information), the background After receiving the user-triggered repository creation message, the information carried in the repository creation message can be used as an input parameter, and the application program interface for creating the repository can be called, thereby using the storage pool specified by the user as the storage in XenServer library.

Step 102: Receive a physical block device PBD creation message triggered by a user, determine all XenServer host information carried in the PBD creation message, create a PBD for connecting each XenServer host and the repository, and mount all the created PBDs.

In the embodiment of the present invention, in order to realize the connection between the established storage warehouse and the XenServer, some or all of the hosts in the XenServer are connected with the storage warehouse by creating a physical block device (PBD).

In practical applications, an application program interface for creating a PBD can be pre-configured. The input parameters of the application program interface are XenServer host information (such as the uuid of each host), repository information (such as the uuid of the repository), and PBD The main function of the attribute information is to create a PBD for connecting the XenServer host and the repository according to the pbd-create command provided by XenServer, and the PBD has the PBD attribute specified by the input parameter.

It should be noted that in the input parameters of the application program interface, there can be one or more XenServer host information, which is determined by the specific implementation of the application program interface. If there can only be one XenServer host information, the repository needs to be established when This step can be performed multiple times when connecting to multiple XenServer hosts.

The following is an example of executing the PBD creation command (pbd-create) provided by XenServer:

#xe pbd-create sr-uuid=4ceb0f8a-1539-40a4-bee2-450a025b04e1host-uuid=83f2c775-57fc-457b-9f98-2b9b0a7dbcb5

In the above command, sr-uuid and host-uuid are the uuid of the repository and the uuid of the host, respectively. After executing the above command, a PBD for connecting the host and the repository can be created (assuming that the uuid of the PBD is aec2c6fc-e1fb-0a27-2437-9862cffe213e, and the uuid of the repository is the one created in the example of step 101 above. uuid of the repository).

In this step, the user can specify the XenServer host, repository, and PBD attribute information in the PBD creation message trigger interface provided by XenServer, and then trigger the PBD creation message (carrying the above specified information). After receiving the user-triggered PBD creation message in the background, The information carried in the PBD creation message can be used as an input parameter, and the application program interface for creating the PBD can be called, thereby realizing the establishment of a PBD for connecting the XenServer host and the storage library specified by the user, and the PBD has the PBD attribute specified by the user. .

In the embodiment of the invention, the PBD attribute information mainly includes the following:

1) Repository installation mode (rbd-mode), mainly has the following values: kernel (this mode is faster but the function is wrong), fuse (this mode is slower, in order to support this mode, rbd-fuse application needs to be installed in XenServer), nbd (this mode is relatively stable In order to support this mode, rbd-nbd application needs to be installed in XenServer). In the present invention, the preferred repository installation mode is the nbd mode, which can be set as a default value.

The implementation code for installing the rbd-fuse application and the rbd-nbd application above is as follows:

2) Driver type (driver-type), there are mainly the following values: vhd, dmp, rbd. In the present invention, the preferred driver type is vhd, which can be set as a default value.

3) Authorized user (cephx-id), indicating the user who uses this PBD, the default value is administrator (admin).

4) Use rbd metadata (use-rbd-meta): used to indicate the storage location of VDI parameters. When the value is True, it indicates that the VDI parameters are stored in the rbd metadata. When the value is False, it indicates that a separate The image stores VDI parameters.

5) Repository prefix (rbd-pool-suffix), used to indicate the name prefix of the repository, such as "RBD_XenStorage".

6) VDI capacity update (vdi-update-existing), which is used to indicate whether the VDI capacity needs to be updated when the storage data of the VDI in the repository decreases. The value is True to indicate that it needs to be updated, and the value is False Indicates that no update is required.

In practical applications, when creating a PBD, if the authorized user of the PBD is admin, it is necessary to judge whether the admin in the distributed storage system and the admin in the XenServer use the same key key file. If so, you can continue to create the PBD. , and set the PBD attribute according to the input parameters of the corresponding application program interface called. Otherwise, the PBD creation cannot be continued. At this time, the PBD creation can be terminated, and the PBD creation failure information can be output at the same time to prompt the user. If the privileged user of the PBD is not admin, you can directly create the PBD and set the properties of the PBD according to the input parameters of the corresponding application program interface called.

In the embodiment of the present invention, after the PBD is created, it is necessary to mount the PBD to truly realize the connection between the XenServer and the repository. Therefore, in the present invention, the PBD can also be mounted according to the PBD mount command (pbd-plug) provided by XenServer in the application program interface for creating PBD, and the application program interface for mounting PBD can also be configured separately. The application program interface for mounting the PBD implements the PBD mounting according to the PBD mounting command provided by the XenServer.

If the application program interface for mounting PBD is configured separately, the PBD mounting can also be triggered by the user. The process is the same as the process of triggering the repository establishment message and PBD creation message above, that is, the user provides the PBD mount message on XenServer. Specify the PBD in the trigger interface, and then trigger the PBD mount message (carrying the above specified PBD information). After receiving the user-triggered PBD mount message in the background, the PBD information carried in the PBD mount message can be used as an input parameter, and the application can be called. The application program interface for mounting the PBD, so as to realize the mounting of the user-specified PBD.

The following is an example of executing the PBD mount command (pbd-create) provided by XenServer:

#xe pbd-plug uuid=aec2c6fc-e1fb-0a27-2437-9862cffe213e

In the above command, the value of uuid (aec2c6fc-e1fb-0a27-2437-9862cffe213e) is the uuid of the PBD to be mounted.

Step 103: Receive a VDI creation message triggered by the user, create a VDI in XenServer and store it in the repository.

In this step, a VDI is created and stored in the repository. Multiple VDIs can be created by performing this step multiple times.

In this embodiment of the present invention, an application program interface for creating VDI may be configured. The input parameters of the application program interface for creating VDI include VDI attribute information (including: the repository to which the VDI belongs, the name of the VDI image, the VDI type, and the size of the VDI image disk). The main functions implemented are: according to the VDI creation command provided by XenServer ( vdi-create) to create a VDI.

In practical applications, the user can specify the VDI attribute information in the VDI creation message trigger interface provided by XenServer, and then trigger the VDI creation message (carrying the VDI attribute information specified by the user). The VDI attribute information carried in the VDI creation message is used as an input parameter, and the application program interface for creating a VDI is invoked to create a VDI.

The following is an example of executing the VDI create command (vdi-create) provided by XenServer:

#xe vdi-create sr-uuid=<sr-uuid>name-label=<image name>type=<system/user>virtual-size=<size_of_virtual_disk file size>.

Step 104: Receive a virtual disk mapping message triggered by the user, and establish a mapping relationship between each VDI in the storage repository and each block device in the distributed storage cluster.

In this step, the block device in the distributed storage cluster is mapped to the VDI in the storage library, and a one-to-one mapping relationship between the block device and the VDI is established, so that the block device in the distributed storage cluster can be used in XenServer. Accessed via mapped VDI.

In this embodiment of the present invention, an application program interface for establishing a mapping relationship between VDI and block devices in a distributed storage cluster may be pre-configured, and the input parameters of the application program interface include repository information (uuid of the repository) and VDI information (uuid of VDI), the main function of the implementation is: according to the mapping command (map) provided by the rbd-nbd application, establish the mapping relationship between a block device in the VDI domain distributed storage cluster in the repository in the input parameter .

In practical applications, the user can specify the repository and VDI in the virtual disk mapping message trigger interface provided by XenServer, and then trigger the virtual disk mapping message (carrying the VDI information and block device information specified by the user), and the background receives the virtual disk mapping message triggered by the user. After the disk mapping message, call the application program interface for establishing the mapping relationship between the VDI and the block device in the distributed storage cluster, and establish a connection between the VDI carried in the virtual disk mapping message and a block device in the distributed storage cluster. Mapping relations.

The following is an example of executing the map command (map) provided by the rbd-nbd application:

rbd-nbd map RBD_XenStorage-cda789de/RBD-17efad10

In the above command, RBD_XenStorage-cda789de before "/" is the uuid of the repository, and RBD-17efad10 after "/" is the uuid of VDI.

This step 104 may be performed multiple times to realize a one-to-one mapping relationship between multiple block devices and multiple VDIs.

It can be seen that the above steps 101 and 102 realize the creation of the repository and connect the repository with XenServer, but no VDI is stored in the repository at this time. The mapping relationship between the block devices in the distributed storage cluster enables the block devices in the distributed storage cluster to be added to XenServer as VDI, thus realizing XenServer's support for distributed storage.

In the embodiment of the present invention shown in FIG. 1, since the distributed storage cluster is a Ceph-based distributed storage cluster, and the block device type in the distributed cluster is rbd, in order to be able to identify the rbd device in XenServer, it is necessary to Configure the rbd block device type in XenServer. Specifically, you can add support for the rbd block device type in the Xapi.conf file of XenServer. To achieve this, you can add the following content to the Xapi.conf file:

In addition, in the above step 101, when the storage pool established in the distributed storage cluster is used as the storage library of XenServer, a message (the value of the device type supported by the storage library carried in the message) needs to be further established based on the storage library. rbd), set the block device type supported by this repository to rbd.

In addition, it is also necessary to modify the tap-ctl file in XenServer, and add the tap disk function for disk format conversion. The modifications are as follows:

Based on the above modifications, when a disk file of the rbd device type is captured in XenServer, the disk format supported by the rbd block device (such as aio format) can be realized by calling the tap disk function to the disk format supported by XenServer (such as vhd format) ) or the opposite on-disk format conversion. Here, when the mapping relationship between the VDI in the repository and the block device in the distributed storage cluster is established, the disk file of the rbd device type can be captured/detected.

In practical applications, after the mapping relationship between the VDI and the block device in the distributed storage cluster is established in the above step 104, the mapping relationship can also be canceled.

An application program interface for canceling the mapping relationship between VDI and block devices in a distributed storage cluster can be pre-configured. The input parameters of the application program interface include repository information (uuid of repository) and VDI information (uuid of VDI), The main function implemented is: according to the unmap command (unmap) provided by the rbd-nbd application program to cancel the mapping relationship of the VDI in the repository in the input parameter (that is, the mapping relationship between the VDI and its corresponding block device).

The user can specify the repository and VDI in the cancel virtual disk mapping message trigger interface provided by XenServer, and then trigger the cancel virtual disk mapping message (carrying the user-specified information). After receiving the user-triggered virtual disk mapping message in the background, call the The application program interface for canceling the mapping relationship between the VDI and the block device in the distributed storage cluster cancels the mapping relationship of the VDI in the storage library carried by the virtual disk mapping message.

The following is an example of executing the map command (map) provided by the rbd-nbd application:

rbd-nbd unmap RBD_XenStorage-cda789de/RBD-17efad10

In the above command, RBD_XenStorage-cda789de before "/" is the uuid of the repository, and RBD-17efad10 after "/" is the uuid of VDI.

In the embodiment of the present invention shown in FIG. 1, operations such as VDI snapshot generation, VDI snapshot data destruction, VDI deletion, and VDI separation are also supported, which are introduced separately below:

Generate a VDI snapshot:

In this embodiment of the present invention, an application program interface for generating a VDI snapshot may be configured. The input parameters of the application program interface for generating VDI snapshots include VDI information (uuid of VDI), and the main functions implemented are: generate snapshots for VDI according to the snapshot generation command (vdi-snapshot) provided by XenServer, and store the snapshot data in The snapshot virtual disk corresponding to the VDI (the snapshot virtual disk corresponding to the VDI can be preset).

In practical applications, the user can specify the VDI in the snapshot generation message trigger interface provided by XenServer, and then trigger the snapshot generation message (carrying the user-specified VDI information). After receiving the user-triggered snapshot generation message in the background, the snapshot can be generated. The VDI information carried in the message is used as an input parameter, and the application program interface for establishing a VDI snapshot is invoked, thereby generating a corresponding snapshot for the VDI.

The following is an example of executing the snapshot generation command (vdi-snapshot) provided by XenServer:

#xe vdi-snapshot uuid=<uuid_of_vdi>.

VDI snapshot data destruction:

In this embodiment of the present invention, an application program interface for destroying the snapshot virtual disk may be configured. The input parameters of the application program interface used to destroy the snapshot virtual disk include VDI information (the uuid of the VDI). .

In practical applications, the user can specify the VDI in the snapshot data destruction message trigger interface provided by XenServer, and then trigger the snapshot data destruction message (carrying the user-specified VDI information). After receiving the user-triggered snapshot data destruction message in the background, you can The VDI information carried in the snapshot data destruction message is used as an input parameter, and the API for destroying the snapshot virtual disk is called, thereby destroying the snapshot virtual disk corresponding to the VDI specified by the snapshot data destruction message.

The following is an example of executing the snapshot data destroy command (vdi-data-destroy) provided by XenServer:

#xe vdi-data-destroy uuid=<uuid_of_vdi>

VDI removal:

In this embodiment of the present invention, an application program interface for deleting VDI may be configured. The input parameters of the application program interface for deleting VDI include VDI information (uuid of VDI), and the main function implemented is to delete the specified VDI according to the VDI delete command (vdi-destroy) provided by XenServer.

In practical applications, the user can specify a VDI in the VDI delete message trigger interface provided by XenServer, and then trigger the VDI delete message (carrying the VDI information specified by the user). After receiving the VDI delete message triggered by the user in the background, the VDI can be deleted. The VDI information carried in the message is used as an input parameter, and the application program interface for deleting VDI is called, thereby deleting the VDI specified by the VDI deletion message.

The following is an example of executing the VDI delete command (vdi-destroy) provided by XenServer:

#xe vdi-data-destroy uuid=<uuid_of_vdi>

VDI separation:

In this embodiment of the present invention, an application program interface for separating VDI may be configured. The input parameters of the API for separating VDI include VDI information (uuid of VDI), and the main function implemented is: separating the specified VDI according to the VDI separation command (vdi-forget) provided by XenServer.

In practical applications, the user can specify the VDI in the VDI separation message trigger interface provided by XenServer, and then trigger the VDI separation message (carrying the VDI information specified by the user). After receiving the user-triggered VDI separation message in the background, the VDI can be separated. The VDI information carried in the message is used as an input parameter, and the application program interface for separating the VDI is called, thereby separating the VDI specified by the VDI separation message.

The following is an example of executing the VDI detach command (vdi-forget) provided by XenServer:

#xe vdi-forget uuid=<uuid_of_vdi>

In the embodiment of the present invention shown in FIG. 1 , disk format conversion during snapshot import/export is also provided.

In this embodiment of the present invention, an application program interface for exporting snapshots may be configured. The input parameters of the API for exporting snapshots include virtual machine information (the uuid of the virtual machine), target file information, and preset disk format. The main functions implemented are: according to the snapshot export command (vm-export) provided by XenServer Exports a snapshot of the specified virtual machine and performs disk format conversion on export.

In practical applications, the user can specify the VM, target file and preset disk format in the snapshot export message trigger interface provided by XenServer, and then trigger the snapshot export message (carrying the information specified by the user), and the user-triggered snapshot export message is received in the background After that, the information carried in the snapshot export message can be used as an input parameter, and the API for exporting snapshots can be called to export the snapshot of the specified VM to the specified target file, and execute the snapshot-supported snapshot of the exported data when exporting. Convert disk format to preset disk format.

The following is an example of executing the snapshot command (vm-export) provided by XenServer:

#xe vm-export vm=0d166725-92a8-f3a7-2817 filename=export.xva

In this embodiment of the present invention, an application program interface for importing snapshots may be configured. The input parameters of the application program interface for importing snapshots include virtual machine information (the uuid of the virtual machine), target file information, and preset disk format. The main functions implemented are: according to the snapshot import command (vm-import) provided by XenServer Imports a snapshot of the specified virtual machine from the specified target file, and performs disk format conversion on import.

In practical applications, the user can specify the VM, target file and preset disk format in the snapshot import message triggering interface provided by XenServer, and then trigger the snapshot import message (carrying the information specified by the user), and receive the user-triggered snapshot import message in the background After that, the information carried in the snapshot import message can be used as an input parameter, and the API for importing snapshots can be called to import the snapshot of the specified VM from the specified target file, and execute the preset disk format to the snapshot during import. Conversion of supported disk formats.

The following is an example of executing the snapshot command (vm-import) provided by XenServer:

#xe vm-import vm=0d166725-92a8-f3a7-2817 filename=export.xva

In actual implementation, the above snapshot import/export and import/export disk format conversion can be implemented by modifying the sparse_dd and vhd-tool files in XenServer, where,

The modifications to sparse_dd are as follows:

The modifications to the vhd-tool file are as follows:

It should be noted that when a VM is created in Xenserver, a disk (ie, a block device) needs to be specified for the VM, and the mapping relationship between the block device and the VDI is established in the present invention. Therefore, the import/ An export is actually an import/export of the data in the VM's disk (ie its corresponding VDI).

In the embodiment of the present invention shown in FIG. 1 , operations such as VDI cloning, VDI expansion, and VDI snapshot merge are also supported, which are introduced separately below:

VDI clone:

In this embodiment of the present invention, an application program interface for cloning VDI may be configured. The input parameters of the application program interface for cloning VDI include VDI information (uuid of VDI), and the main function implemented is: clone the specified VDI according to the VDI clone command (vdi-clone) provided by XenServer.

In practical applications, the user can specify the VDI in the VDI clone message trigger interface provided by XenServer, and then trigger the VDI clone message (carrying the user-specified VDI information). After receiving the user-triggered VDI clone message in the background, the VDI clone can be cloned. The VDI information carried in the message is used as an input parameter, and the API for cloning the VDI is called to clone the VDI specified by the VDI clone message.

The following is an example of executing the VDI clone command (vdi-clone) provided by XenServer:

#xe vdi-clone uuid=<uuid_of_the_vdi>

VDI expansion:

In this embodiment of the present invention, an application program interface for VDI expansion can be configured. The input parameters of the application program interface used for VDI expansion include VDI information (uuid of VDI) and expansion information. The main function implemented is to expand the specified VDI according to the VDI expansion command (vdi-resize) provided by XenServer.

In practical applications, the user can specify VDI and expansion information (size after expansion) in the VDI expansion message trigger interface provided by XenServer, and then trigger the VDI expansion message (carrying the VDI information and expansion information specified by the user), and the background receives the user trigger After the VDI expansion message is sent, the VDI information and expansion information carried in the VDI expansion message can be used as input parameters, and the application program interface for VDI expansion can be called to expand the specified VDI according to the expansion information specified in the VDI expansion message. .

The following is an example of executing the VDI resize command (vdi-resize) provided by XenServer:

#xe vdi-resize uuid=<vdi_uuid>disk-size=<new_size_for_disk>

VDI snapback:

In this embodiment of the present invention, an application program interface for merging VDI snapshots may be configured. The input parameters of the application program interface for merging VDI snapshots include snapshot information (snapshot uuid) and target file information. The main function implemented is: according to the snapshot merge command (snapshot-export-to-template) provided by XenServer Snapshots are merged into the target file.

In practical applications, the user can specify the snapshot and target file in the snapshot merge message trigger interface provided by XenServer, and then trigger the snapshot merge message (carrying the snapshot information and target file information specified by the user), and receive the user-triggered snapshot merge message in the background. Afterwards, the snapshot information and target file information carried in the snapshot merging message can be used as input parameters, and the application program interface for merging VDI snapshots can be called, thereby merging the snapshot specified in the snapshot merging message into the target file.

The following is an example of executing the snapback command (snapshot-export-to-template) provided by XenServer:

#xe snapshot-export-to-template snapshot-uuid=30e4b27d-2e95-8a87-c340 filename=/win07bak.xva

The method for realizing distributed storage in XenServer of the present invention is described in detail above. The present invention also provides a plug-in for realizing distributed storage in XenServer, which is described in detail below with reference to FIG. 2 .

Referring to FIG. 2, FIG. 2 is a schematic structural diagram of a plug-in implementing distributed storage in XenServer according to an embodiment of the present invention. As shown in FIG. 2, the plug-in includes: a receiving unit 201 and a processing unit 202; wherein,

The receiving unit 201 is configured to receive a user-triggered repository establishment message; be used to receive a user-triggered physical block device PBD creation message; be used to receive a user-triggered virtual disk image VDI creation message; be used to receive a user-triggered virtual disk mapping information;

The processing unit 202 is used to use the storage pool established in the distributed storage cluster as the storage library of XenServer when the receiving unit 201 receives the user-triggered repository establishment message; for the receiving unit 201 to receive the user-triggered PBD creation During the message, determine all the XenServer host information carried by the PBD creation message, create a PBD for connecting each XenServer host and the storage library, and mount all the created PBDs; for the receiving unit 201 to receive the user-triggered VDI creation When the message is sent, a VDI is created in XenServer and stored in the repository; when the receiving unit 201 receives a user-triggered virtual disk mapping message, it creates a relationship between each block device in the distributed storage cluster and each VDI in the repository. Mapping relations.

The plug-in shown in FIG. 2 also includes a configuration unit 203;

The distributed storage cluster is a Ceph-based distributed storage cluster, and the block device type in the distributed cluster is rbd;

The configuration unit 203 is used to configure the rbd block device type in XenServer in advance;

The processing unit 202, when using the storage pool established in the distributed storage cluster as the storage library of XenServer, further establishes a message based on the storage library, and sets the block device type supported by the storage library to rbd;

The processing unit 202, after establishing the mapping relationship between each block device in the distributed storage cluster and each VDI in the storage library, further includes: when a block device of type rbd is captured, changing the disk format of the block device from the The disk format supported by the distributed storage cluster is converted to the disk format supported by XenServer.

In the plug-in shown in Figure 2,

The PBD creation message carries PBD attribute information, and the PBD attribute information includes authority user information and repository installation mode;

The processing unit 202, when creating a PBD for connecting each XenServer host and the repository, and when mounting all the created PBDs, further determines that the PBD creation message carries the authority user information in the PBD attribute information, if the If the authorized user is an administrator, then determine whether the administrator of the distributed storage cluster and the administrator use the same key file. If so, configure the properties of the created PBD according to the PBD property information. If not, terminate the PBD creation , and output PBD creation failure information; if the authorized user is not an administrator, configure the properties of the created PBD according to the PBD property information.

In the plug-in shown in Figure 2,

The receiving unit 201 is further configured to receive a user-triggered virtual disk mapping cancellation message;

The processing unit 202 is further configured to, when the receiving unit 201 receives a user-triggered mapping cancellation message, cancel the rbd block device based on the mapping cancellation command provided by the Xenserver according to the rbd fast device and VDI information carried in the mapping cancellation message The mapping relationship with this VDI.

In the plug-in shown in Figure 2,

The receiving unit 201 is further configured to receive a user-triggered snapshot generation message, a user-triggered snapshot data destruction message, a user-triggered VDI deletion message, and a user-triggered VDI separation message;

The processing unit 202 is further configured to, when the receiving unit 201 receives a user-triggered snapshot generation message, create a snapshot corresponding to the VDI specified by the snapshot generation message based on the snapshot generation command provided by the Xenserver, and store the snapshot data in the snapshot generation message. The snapshot virtual disk corresponding to the VDI; when the receiving unit 201 receives the snapshot data destruction message triggered by the user, based on the snapshot data destruction command provided by Xenserver, the snapshot virtual disk corresponding to the VDI specified by the snapshot data destruction message is destroyed; in the receiving unit When receiving the VDI deletion message triggered by the user, 201 deletes the VDI specified by the VDI deletion message based on the VDI deletion command provided by Xenserver; when the receiving unit 201 receives the VDI separation message triggered by the user, based on the VDI separation command provided by Xenserver , detach the VDI specified in the VDI detach message.

In the plug-in shown in Figure 2,

The receiving unit 201 is further configured to receive a user-triggered snapshot export message; configured to receive a user-triggered snapshot import message;

The processing unit 202 is further configured to, when the receiving unit 201 receives the snapshot export message triggered by the user, based on the snapshot export command provided by XenServer, export the snapshot of the virtual machine vm specified by the snapshot export message to the snapshot specified by the snapshot export message. target file, and perform the conversion of the disk format supported by the snapshot to the preset disk format when exporting; for when the receiving unit 201 receives the snapshot import message triggered by the user, based on the snapshot import command provided by XenServer, the snapshot is imported The snapshot of the virtual machine vm specified in the message is imported from the target file specified in the snapshot import message, and the conversion from the preset disk format to the disk format supported by the snapshot is performed during import.

In the plug-in shown in Figure 2,

The receiving unit 201 is further configured to: receive a user-triggered VDI clone message, receive a user-triggered VDI capacity expansion message, and receive a user-triggered VDI capacity expansion message;

The processing unit 202 is further configured to clone the VDI specified by the VDI clone message based on the VDI clone command provided by the Xenserver when the receiving unit 201 receives the VDI clone message triggered by the user; During the VDI expansion message, determine the VDI information and the expansion information carried in the VDI expansion message, and based on the VDI expansion command provided by Xenserver, expand the VDI according to the expansion information; when the receiving unit 201 receives the user-triggered snapshot merge message , determine the snapshot information and target file information carried in the snapshot merge message, and merge the snapshot into the target file based on the snapshot merge command provided by Xenserver.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the present invention. within the scope of protection.

Claims (14)

1. A method for implementing distributed storage in a XenServer, the method comprising:

receiving a storage library establishing message triggered by a user, and using a storage pool established in the distributed storage cluster as a storage library of the XenServer;

receiving a Physical Block Device (PBD) creation message triggered by a user, determining all XenServer host information carried by the PBD creation message, creating a PBD for connecting each XenServer host with the storage library, and mounting all the created PBDs;

receiving a Virtual Disk Image (VDI) establishing message triggered by a user, establishing a VDI in a XenServer and storing the VDI in the storage library;

and receiving a virtual disk mapping message triggered by a user, and establishing a mapping relation between each piece of equipment in the distributed storage cluster and each VDI in the storage library.

2. The method of claim 1,

the distributed storage cluster is a distributed storage cluster based on Ceph, and the type of block equipment in the distributed storage cluster is rbd;

configuring the type of the rbd block device in the XenServer in advance;

when the storage pool established in the distributed storage cluster is used as a storage pool of the XenServer, the type of the block device supported by the storage pool is set to be rbd further based on the storage pool establishment message;

after the mapping relationship between each piece of device in the distributed storage cluster and each VDI in the repository is established, the method further includes: and when the rbd type block device is captured, converting the disk format of the block device from the disk format supported by the distributed storage cluster into the disk format supported by the XenServer.

3. The method of claim 2,

the PBD establishing message carries PBD attribute information, and the PBD attribute information comprises authority user information and a memory bank installation mode;

when a PBD used for connecting each XenServer host and the memory bank is created, and all created PBDs are mounted, further determining that the PBD creation message carries authority user information in PBD attribute information, if the authority user is an administrator, judging whether the administrator of the distributed storage cluster and the administrator use the same key file, if so, performing attribute configuration on the created PBD according to the PBD attribute information, if not, terminating the PBD creation, and outputting PBD creation failure information; and if the authority user is not the administrator, performing attribute configuration on the created PBD according to the PBD attribute information.

4. The method of claim 2, further comprising:

receiving a virtual disk mapping canceling message triggered by a user, determining the information of the rbd fast device and the VDI carried by the virtual disk mapping canceling message, and canceling the mapping relation between the rbd block device and the VDI based on a mapping canceling command provided by the Xenerver.

5. The method of claim 2, further comprising:

receiving a snapshot generating message triggered by a user, establishing a snapshot corresponding to a VDI (virtual desktop infrastructure) specified by the snapshot generating message based on a snapshot generating command provided by the Xenerver, and storing snapshot data in a snapshot virtual disk corresponding to the VDI;

receiving a snapshot data destruction message triggered by a user, and destroying a snapshot virtual disk corresponding to a VDI (virtual desktop infrastructure) specified by the snapshot data destruction message based on a snapshot data destruction command provided by the Xenerver;

receiving a VDI deletion message triggered by a user, and deleting the VDI specified by the VDI deletion message based on a VDI deletion command provided by the Xenerver;

and receiving a VDI separation message triggered by a user, and separating the VDI specified by the VDI separation message based on a VDI separation command provided by the Xenerver.

6. The method of claim 5, further comprising:

receiving a snapshot export message triggered by a user, exporting a snapshot of a virtual machine vm specified by the snapshot export message to a target file specified by the snapshot export message based on a snapshot export command provided by a XenServer, and executing conversion from a disk format supported by the snapshot to a preset disk format during export;

receiving a snapshot import message triggered by a user, importing the snapshot of the virtual machine vm specified by the snapshot import message from a target file specified by the snapshot import message based on a snapshot import command provided by the XenServer, and executing conversion from a preset disk format to a disk format supported by the snapshot during import.

7. The method of claim 2, further comprising:

receiving a VDI cloning message triggered by a user, and cloning a VDI specified by the VDI cloning message based on a VDI cloning command provided by the Xenerver;

receiving a VDI capacity expansion message triggered by a user, determining VDI information and capacity expansion information carried by the VDI capacity expansion message, and based on a VDI capacity expansion command provided by Xenerver, carrying out capacity expansion on the VDI according to the capacity expansion information;

receiving a snapshot merging message triggered by a user, determining snapshot information and target file information carried by the snapshot merging message, and merging the snapshot into the target file based on a snapshot merging command provided by the Xenerver.

8. A plugin for implementing distributed storage in a XenServer, the plugin comprising: a receiving unit and a processing unit;

the receiving unit is used for receiving a repository establishment message triggered by a user; the PBD creation message is used for receiving a user-triggered PBD creation message; the virtual disk image management system is used for receiving a virtual disk image VDI creating message triggered by a user; the virtual disk mapping message is used for receiving a virtual disk mapping message triggered by a user;

the processing unit is used for using the storage pool established in the distributed storage cluster as a storage pool of the XenServer when the receiving unit receives the storage pool establishing message triggered by the user; the receiving unit is used for determining all XenServer host information carried by the PBD creation message when receiving the PBD creation message triggered by the user, creating the PBD for connecting each XenServer host with the storage library and mounting all the created PBDs; the receiving unit is used for creating the VDI in the XenServer and storing the VDI in the storage library when receiving a VDI creation message triggered by a user; the receiving unit is used for establishing the mapping relation between each piece of equipment in the distributed storage cluster and each VDI in the storage library when receiving a virtual disk mapping message triggered by a user.

9. An insert as claimed in claim 8, further comprising a configuration unit;

the distributed storage cluster is a distributed storage cluster based on Ceph, and the type of block equipment in the distributed storage cluster is rbd;

the configuration unit is used for configuring the type of the rbd block device in the XenServer in advance;

when the storage pool established in the distributed storage cluster is used as a storage pool of the XenServer, the processing unit further sets the block device type supported by the storage pool to be rbd based on the storage pool establishment message;

after the processing unit establishes the mapping relationship between each piece of device in the distributed storage cluster and each VDI in the repository, the processing unit further includes: and when the rbd type block device is captured, converting the disk format of the block device from the disk format supported by the distributed storage cluster into the disk format supported by the XenServer.

10. The insert of claim 9,

the PBD establishing message carries PBD attribute information, and the PBD attribute information comprises authority user information and a memory bank installation mode;

the processing unit is used for creating a PBD for connecting each XenServer host with the memory bank, further determining that the PBD creation message carries authority user information in PBD attribute information when mounting all the created PBDs, judging whether an administrator of the distributed storage cluster and the administrator use the same key file if the authority user is the administrator, performing attribute configuration on the created PBD according to the PBD attribute information if the authority user is the administrator, and stopping PBD creation and outputting PBD creation failure information if the authority user is the administrator; and if the authority user is not the administrator, performing attribute configuration on the created PBD according to the PBD attribute information.

11. The insert of claim 9,

the receiving unit is further configured to receive a virtual disk mapping cancellation message triggered by a user;

the processing unit is further configured to, when the receiving unit receives a mapping canceling message triggered by a user, cancel a mapping relationship between the rbd block device and the VDI based on a mapping canceling command provided by the Xenserver according to the rbd fast device and the VDI information carried in the mapping canceling message.

12. The insert of claim 9,

the receiving unit is further configured to receive a snapshot generation message triggered by a user, a snapshot data destruction message triggered by the user, a VDI deletion message triggered by the user, and a VDI separation message triggered by the user;

the processing unit is further configured to, when the receiving unit receives a snapshot generating message triggered by a user, establish a snapshot corresponding to the VDI specified by the snapshot generating message based on a snapshot generating command provided by the Xenserver, and store snapshot data in a snapshot virtual disk corresponding to the VDI; when a receiving unit receives a snapshot data destruction message triggered by a user, destroying a snapshot virtual disk corresponding to a VDI (virtual desktop infrastructure) specified by the snapshot data destruction message based on a snapshot data destruction command provided by the Xenserver; when a receiving unit receives a VDI deletion message triggered by a user, deleting the VDI specified by the VDI deletion message based on a VDI deletion command provided by the Xenerver; when the receiving unit receives a VDI separation message triggered by a user, the VDI specified by the VDI separation message is separated based on a VDI separation command provided by the Xenerver.

13. The insert of claim 12,

the receiving unit is further configured to receive a snapshot export message triggered by a user; the snapshot import message is used for receiving a snapshot triggered by a user;

the processing unit is further configured to, when the receiving unit receives a snapshot export message triggered by a user, export a snapshot of the virtual machine vm specified by the snapshot export message to a target file specified by the snapshot export message based on a snapshot export command provided by the XenServer, and perform conversion from a disk format supported by the snapshot to a preset disk format when exporting the snapshot export command; the method and the device are used for importing the snapshot of the virtual machine vm specified by the snapshot import message from the target file specified by the snapshot import message based on the snapshot import command provided by the XenServer when the receiving unit receives the snapshot import message triggered by the user, and executing the conversion from the preset disk format to the disk format supported by the snapshot during the import.

14. The insert of claim 9,

the receiving unit is further configured to: receiving a VDI cloning message triggered by a user, receiving a VDI capacity expansion message triggered by the user and receiving a VDI capacity expansion message triggered by the user;

the processing unit is further configured to clone the VDI specified by the VDI clone message based on a VDI clone command provided by the Xenserver when the receiving unit receives the VDI clone message triggered by the user; when a receiving unit receives a VDI capacity expansion message triggered by a user, VDI information and capacity expansion information carried by the VDI capacity expansion message are determined, and the VDI is subjected to capacity expansion according to the capacity expansion information based on a VDI capacity expansion command provided by Xenerver; when a receiving unit receives a snapshot merging message triggered by a user, snapshot information and target file information carried by the snapshot merging message are determined, and the snapshot is merged into the target file based on a snapshot merging command provided by the Xenerver.

Images