CN103744625A - Method for copying disk data in virtual application - Google Patents

Abstract

The invention provides a method for copying disk data in a virtual application. According to the method, load of a server host is transferred to a storage hardware system when a virtual machine requires copying and cloning a virtual disk remotely, the storage hardware system performs IO operations according to dynamic characteristics of the request during the process, and multisource multi-target copying is allowed, so that system performance is greatly improved. Compared with the prior art, the method has the advantages that through the multisource multi-target copying in the storage virtual application, copying disorder caused by overlapping of copying source areas and copying target areas is avoided; for read-only files, mapping of the copying target areas is disconnected to achieve copying effect, so that IO response speed of the storage system is increased, and market competitiveness of company products is improved.

Description

In a kind of virtualization applications, copy the method for data in magnetic disk

Technical field

The present invention relates to computer virtualized applied technical field, more specifically say the method that copies data in magnetic disk in virtualization applications.

Background technology

The architecture of VMware virtual platform based on can input business using builds.Use the softwares such as VMware vSphere to change or the hardware resource (comprising CPU, RAM, hard disk and network controller) of " virtual " computing machine based on x86, virtual machine multiple functional to create, can move himself operating system and application program as " truly " computing machine.Each virtual machine comprises a set of complete system, thereby does not have potential conflict.The virtualized principle of work of VMware is directly at computer hardware or above host operating system, to insert a software layer of simplifying.This software layer comprises a virtual machine monitor (i.e. " virtual management program ") that distributes hardware resource with dynamic and transparent mode.Multiple operating system can operate in separate unit physical machine simultaneously, shares each other hardware resource.Because being is encapsulated whole computing machine (comprising CPU, internal memory, operating system and the network equipment), therefore virtual machine can be completely compatible with x86 operating system, application program and the device driver of all standards.The multiple operating system of safe operation and application program in single computer simultaneously, each operating system and application program can be accessed its required resource when needed.

ESX is the enterprise-level virtualization product of VMware, is the virtualized platform base of vmware, is deployed on server.It is that to have advanced resource management function efficient, fictitious host computer platform flexibly, perfect matching enterprise data center, by improving resource utilization, expansion computing power and Optimized Service device, VMware ESX server helps enterprise to reduce the cost of computer infrastructure.

From the angle of broad sense, vStorage API is a kind of mechanism that can allow virtual foundation suite and storage hardware more combine closely:

For virtual external member provides the storage hardware of solution.

VMware or third party's management software.

For the application program operating in virtual machine provides better store quality of service.

VAAI(vStorage API for Array Integration) be a subset of the vStorage API standard that scope is larger, be mainly to utilize hardware offload characteristic improving performance.

In virtualization applications, when system is carried out remote copy and clone's virtual disk, if use traditional clone method, will be very consuming time, inefficiency.In order to solve existing this generic operation, to server host, bring resource consumption excessive, the problem of overlong time, realization transfers load to storage hardware, Hoisting System overall performance, a kind of method entirely copying that is called that improves data in magnetic disk copy efficiency in Storage Virtualization application is now provided, and the method sends full copy command request to storage system during by remote copy, will greatly improve copy efficiency, Hoisting System overall performance, alleviates server load.

Summary of the invention

Technical assignment of the present invention is to solve the deficiencies in the prior art, and a kind of method that copies data in magnetic disk in virtualization applications is provided.

Technical scheme of the present invention realizes in the following manner, copies the method for data in magnetic disk in this kind of virtualization applications, and its specific implementation process is:

One, storage array hardware is received the initialization requests from fictitious host computer, and distribute required resource, and check its legitimacy, wherein validity checking comprises task status inspection and parameter validity checking, if illegally will stop this request, and return to specific perception data;

Two, execution parameter legitimacy detects, and whether if asynchronous mode request, needing to detect has identical order to process;

Three, to copy source and target LUN, send querying command, according to returning results, determine whether LUN can normally access;

Four, judge that whether copy source and target region is overlapping;

Five, carry out the data copy of overlapping region, and mark;

Six, order executing data copy;

Seven, retain the status data of command processing result and order self reservation data;

Eight, to server, return to result;

Nine, according to timer, judge whether retain data time expires, if not yet due, continue to retain data, if expired, finish whole process.

The beneficial effect that the present invention compared with prior art produced is:

The method that copies data in magnetic disk in a kind of virtualization applications of the present invention adopts asks remote copy and the load of clone's virtual disk to be transferred in storage hardware system the virtual machine in server host, in this process, storage system is according to the feature Dynamic Execution IO operation of this request, allow how far multiple goal copy function, greatly improve system performance; Storage system can be chosen in the upper bob that disconnects copy targeting of carrying out from the LUN simplifying, and with copy source replacement copy targeting, real diskcopy operation, but postpones till when write operation is carried out in this data area; Can prevent because of the overlapping copy confusion causing in behaviour region, copy source and target area; Improve wrong treatment mechanism, retain done state data and reservation data that request is carried out, allow user that retention time is set voluntarily; Solve existing this generic operation and bring resource consumption excessive to server host, the problem of overlong time, realizes and transfers load to storage hardware, Hoisting System overall performance; Practical, be easy to promote.

Accompanying drawing explanation

Accompanying drawing 1 is the static diagram on full reproduction process virtual disk in virtualization applications of the present invention.

Accompanying drawing 2 is full copy data copy process process flow diagrams of the present invention.

Embodiment

Below in conjunction with accompanying drawing, the method that copies data in magnetic disk in a kind of virtualization applications of the present invention is described in detail below.

The core concept that system involved in the present invention realizes is: the load that the full copy command in virtualized VAAI characteristic can alleviate server host end, and storage array hardware is in the process that realizes this unloading, can be according to the dynamic process that requires of host side, actual copy can not be carried out in storage, and just carry out a process remapping, can greatly improve the speed of copy procedure; If it is overlapping that the duplication source data region in copy procedure and target data region have, first copy overlapping region, and then order copies remaining data.Will prevent like this copies data because confusion overlapping and that cause; Storage system can be supported synchronization request and two kinds of patterns of Asynchronous Request of entirely copying; Due to full reproduction process complexity, occur that wrong probability is large, so need to improve error mechanism processing.Full reproduction process will retain command execution done state data and retain data, and its retention time length is set by user, can satisfy the demand like this.

As shown in accompanying drawing 1, Fig. 2, now provide a kind of method that copies data in magnetic disk in virtualization applications.The static diagram entirely copying in virtualization applications is provided in Fig. 1, and disk is mapped as file through virtual level VMFS file system, and the raw device on the right forms contrast.The file being present on LUN A in VMFS file system is cloned on LUN B, result source file section <s1, e2> and <s4, e1> is cloned into file destination section <s4, e3> and <s1, eN>, and storage system can not carried out real copy function, but be only to disconnect file destination section <s4, e3> and <s1, eN>, with source file section <s1, e2> and <s4, e1> replaces, and be read-only by this file mark, when needs are carried out write operation to this file, carry out again copy work.

Accompanying drawing 2 has been described the concrete treatment scheme of initialization, mainly contains following flow process and realizes:

Flow process 100: storage array hardware is received the initialization requests from fictitious host computer, and distribute required resource, check its legitimacy.Wherein validity checking comprises task status inspection and parameter validity checking, if illegally will stop this request, and returns to specific perception data.

Flow process 101: execution parameter legitimacy detects, whether if asynchronous mode request, needing to detect has identical order to process.

Flow process 102: LUN sends INRUIRY querying command to copy source and target, determines according to returning results whether LUN can normally access.

Flow process 103: judge that whether copy source and target region is overlapping.

Flow process 104: carry out the data copy of overlapping region, and mark.

Flow process 105: order executing data copy.

Flow process 106: retain the status data of command processing result and the reservation data of order self.

Flow process 107: return to result to server.

Flow process 108: judge according to timer whether retain data time expires.

Flow process 109: continue to retain data.

In the present invention, in executing data copy procedure, particularly have in the situation of multiple source and targets, server no longer needs repeatedly to storage system, to send the order of copies data, but only sends a subcommand to storage system.Then this order of storage system asynchronous process, processing finish or abnormal end situation under to server, return to result, server is made corresponding operation according to this result.The method or the IO load that greatly alleviates server, optimized system.

At some in particular cases, while carrying out copy in same LUN, copy source and copy object region overlap.If use traditional data copying method, copies data may be destroyed source data because of covering, causes copy procedure confusion.In the method, can first detect overlapping duplication source data, then, first by carrying out the data copy of this overlapping region, then carry out normal copy function.So just avoided because the chaotic generation of copy.

The applied environment occurring in copies data process is very complicated, and the type making a mistake is also many, and existing copy method mistake is processed comprehensive not.After execution has copied, storage system need to be that the result of this operation and the data of reservation retain a period of time, if exceed client during this period of time, sends these data of inquiry, can point out mistake.In the method, will client be returned to more friendly reminder-data, and allow client to revise the time that retains data.

Scheme provided by the invention, by IO request and system status are assessed, well Hoisting System performance.

Certainly; the present invention also can have other various embodiments; in the situation that not deviating from spirit of the present invention and essence thereof; those of ordinary skill in the art can make according to the present invention various corresponding changes and distortion, but these corresponding changes and distortion all should belong to the protection domain of the appended claim of the present invention.

One of ordinary skill in the art will appreciate that all or part of step in said method can carry out instruction related hardware by program and complete, described program can be stored in computer-readable recording medium, as ROM (read-only memory), disk or CD etc.Alternatively, all or part of step of above-described embodiment also can realize with one or more integrated circuit.Correspondingly, the each module/unit in above-described embodiment can adopt the form of hardware to realize, and also can adopt the form of software function module to realize.The present invention is not restricted to the combination of the hardware and software of any particular form.

The foregoing is only embodiments of the invention, within the spirit and principles in the present invention all, any modification of doing, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.

Claims (1)

1. in virtualization applications, copy a method for data in magnetic disk, it is characterized in that its specific implementation process is:

One, storage array hardware is received the initialization requests from fictitious host computer, and distribute required resource, and check its legitimacy, wherein validity checking comprises task status inspection and parameter validity checking, if illegally will stop this request, and return to specific perception data;

Two, execution parameter legitimacy detects, and whether if asynchronous mode request, needing to detect has identical order to process;

Three, to copy source and target LUN, send querying command, according to returning results, determine whether LUN can normally access;

Four, judge that whether copy source and target region is overlapping;

Five, carry out the data copy of overlapping region, and mark;

Six, order executing data copy;

Seven, retain the status data of command processing result and order self reservation data;

Eight, to server, return to result;

Nine, according to timer, judge whether retain data time expires, if not yet due, continue to retain data, if expired, finish whole process.

Images