KR101673882B1 - Storage system with virtualization using embedded disk and method of operation thereof - Google Patents

Abstract

The present invention relates to a method for providing high availability using the embedded disk of a host and a virtualization based storage system for performing the method. The virtualization based storage system includes a virtualization and cloud solution unit for registering information on a recognized integrated mount point and connecting an external storage, a storage high availability manager for performing clustering, replication/distribution, and problem process, and an external storage part for supporting the high availability of a storage. So, the local disk of the physical server can function as the external storage.

Description

TECHNICAL FIELD [0001] The present invention relates to a high availability virtualization-based storage system using a built-in disk and a storage system using the same,

The present invention relates to a server-internal disk-based external storage provisioning technology for a highly available virtual machine service on a cloud, and more particularly, to a disk-based external storage provisioning system for creating block storage for a virtual machine (VM) in a cloud or virtualization environment, The present invention relates to a method for providing high availability using a built-in disk of a host which enables storage and management using the storage, and a virtualization-based storage system performing the method.

Traditional IT infrastructure required multiple server, storage, and networking equipment to provide a single IT service. However, it is necessary to introduce such a large number of servers and storage for one service, and as the number of kinds of services increases, the number of equipment to be introduced increases.

First, there are many difficulties in management. There are many difficulties to manage because it is necessary to know what equipment is running in order to provide each IT service, know where each equipment is located, and what applications and data are stored in each equipment.

Second, we need to worry about multiplexing configuration for stable operation. In other words, if the system is operated only as a single unit, it will result in service interruption in case of failure. Therefore, in order to solve this problem, it is necessary to configure multiple units to overcome the obstacles that may occur. Larger equipment needs to be introduced.

Third, a lot of resources are inevitably consumed because more space and more power are consumed by introducing and configuring a large amount of servers and storage. Therefore, IT infrastructure operation costs a lot. In order to solve these problems, we improved the introduction method of existing service unit, virtualized server and storage by using virtualization technology and cloud technology, and used infrastructure to manage IT infrastructure by software, By reducing the size of the introduction, it is possible to save physical space and electricity bills.

However, even under such virtualization-based IT infrastructure operating environment, if a physical virtualization server running a virtual machine service fails, the operation of the virtual machine running in the virtualization server is interrupted.

The only way to solve this problem is to use external storage as a way to utilize virtualization technology. In order to maintain stable and non-disruptive high availability of virtual machine services, And the virtual machine data is externally operated so that the system can be configured to cope with the failure by using the virtual machine data stored in the outside even when the failure occurs. In other words, external storage should be separately configured in addition to the server to improve reliability. However, compared to traditional IT infrastructures, it has been able to reduce the size of server adoption, resulting in many improvements in management and cost. However, the introduction of external storage is essential for stable virtualization services and does not fundamentally solve the problem .

The background art of the present invention is disclosed in Korean Patent Publication No. 10-2013-000004832 (2013.01.14) filed in the Korean Intellectual Property Office.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and it is an object of the present invention to provide a method and apparatus for connecting a virtual machine, which is operated in a virtualized IT infrastructure environment, to an external storage in order to overcome a failure of a host, The objective is to solve the structural problem and replace the role of external storage by using the local disk of the physical server.

Another object of the present invention is to provide a storage controller function having a data distribution and replication function in order to enhance the stability of a virtual machine in achieving the above object, thereby enabling stable storage management and high availability virtual machine service, And provide high-performance virtualization-based cloud systems.

In order to provide high availability of a virtual machine according to an exemplary embodiment of the present invention, a virtualization-based storage system using a built-in disk in a server registers information on the recognized integrated mount point to provide virtualization and clutter The solution includes clustering, replication and distribution for storage high availability, a storage high availability manager responsible for failover, and external storage for high availability of storage.

The storage high availability manager includes a storage cluster manager unit for clustering external storage of a host, an external storage unit for creating an integrated mount point so that the clustered external storage can be mounted as one unit, A storage cloning and distribution manager for deciding whether to copy, distribute, or divide the data, and a high availability processing manager unit for constantly replicating the same data to a host having no failure when the host fails.

The external storage providing unit not only provides an integrated mount point for clustering external storages of the respective hosts through the storage cluster manager unit and allowing the clustered external storage to be mounted as one, A volume mount unit for determining which disk is to be mounted in which volume, and an external access point unit for determining which external volume is to be recognized and connected to the mounted volume.

A disk-based external storage system for a high availability virtual machine service can be configured to provide a full virtualization service or a cloud service only with a virtualization server without an external storage. Therefore, a virtualization service can be configured with a simple structure without a complicated structure There is an advantage that it can be.

In addition, because the configuration is simple and simple, the cost of configuring virtualization services or cloud services is low, and the overall TCO is lowered as the space and power consumption of the devices is reduced and management efficiency is improved.

In addition, it supports replication and decentralized structure internally and improves the stability of data and stability of virtualization service or cloud service due to failure handling mechanism to overcome automatically in case of failure, as well as enhancing data safety by preventing loss of data. .

FIG. 1 schematically illustrates a structure of a disk-based external storage system with a built-in server for a high-availability fusing machine service according to an embodiment of the present invention.
FIG. 2 illustrates operations when a failure occurs in a server-based disk-based external storage system for a high availability fusing machine service according to an exemplary embodiment of the present invention.
FIG. 3 is a conceptual diagram of an internal storage-based external storage providing technology according to an embodiment of the present invention.
4 illustrates a concept of an internal data replication and distribution structure of a clustered disk according to an embodiment of the present invention.
FIG. 5 illustrates a flow of a method of virtualizing using a built-in disk in a server according to an embodiment of the present invention.
6 shows a flowchart of a clustering method of a built-in disk in a server according to an embodiment of the present invention.
7 illustrates a conceptual diagram of an internal data replication / distribution structure of a clustered disk according to an embodiment of the present invention
FIG. 8 is a conceptual diagram illustrating failure handling and failure recovery of a disk according to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 and FIG. 2 are views illustrating a structure of a disk-based external storage system built into a server for a high availability fusing machine service according to an embodiment of the present invention.

Referring to FIGS. 1 and 2, the external storage system 100 includes a plurality of hosts. In FIG. 1 of the present invention, two hosts are illustrated. The external storage system 100 includes a virtualization / cloud solution unit 110 for connecting external storage, a storage high availability manager unit 120 for clustering, replication / distribution, fault handling, and high availability of storage And an external storage unit 130 for storing the external storage unit.

When a failure occurs in the host 1, the storage high availability manager unit 120 of the host 2 maintains the mount point 230 so that the external storage can be used regardless of the failure of the host 1. The virtualization / cloud solution unit 110 automatically transfers the virtual machines (VMs) running on the host 1 to the host 2 to quickly overcome the failure. The storage high availability manager 120 performs clustering, duplication / distribution, and failure handling of the external storage using the detailed modules.

The storage high availability manager unit 120 includes a storage cluster manager unit 122 for clustering external storage, an external storage provision unit 121 for creating an integrated mount point 320 so that the clustered external storage can be mounted as one, A storage cloning / distribution manager unit 123 for determining whether to copy / distribute / divide data to / from the internal disks in the host, a host which does not fail when the host fails, And a high availability processing manager unit 124.

The external storage provider 121 clusters the external storage of each host through the storage clustering manager unit 123 and provides an integrated mount point 320 so that the clustered external storage can be mounted as one.

The external storage unit 121 includes a volume mount unit 220 for determining which disk is to be mounted on the internal disk so that the internal disk can be recognized by the server, a volume mount unit 220 for recognizing the mounted volume as an external volume, And a mount point 230 for connecting the external access point unit 210 to the external storage and the external access point unit, as shown in FIG.

When the server is operated, the external storage providing unit 121 activates the transducer in the external storage unit 121 to recognize the external storage of the internal disk, which is not shown in FIG. 3 of the present invention. The converter checks the type of internal disk built in the server, the RAID configuration, and the total size of the disk. If the internal disk type is wrong or the RAID configuration is wrong, inform the user to convert it to the correct disk type, Guide. The converter mounts the disk so that the server recognizes the host's internal disk and decides which disk to mount to which volume. The mounted disk is recognized as the host's local volume and is available only to that host. The converter determines which external volume the mounted volume is to be recognized as. Each volume is made into a different external volume, and the external storage unit 121 is activated to start the storage service. The external storage unit 121 has an IP address such as general external storage and can be connected to the storage unit through an IP address.

Figure 4 shows the concept of an internal data replication and distribution structure of a clustered disk. The present invention provides clustering technology of a host-mounted disk that supports recognizing internal disks in the virtualization and cloud service host server as one storage and expanding capacity.

In FIG. 4, each host has an address of the external storage recognized through the external storage provision technology for the built-in storage of the host provided earlier. And uses the IP address of each storage to recognize that each server is the IP address of an external storage used for a single cloud or virtualization service. If you convert the internal disks of four hosts to external storage, you have four external storage IPs. And clustering technology to register information so that four external storage devices are used for a single cloud or virtualization service. Each recognized external storage uses technology to be recognized as a single storage through the storage cluster manager 122. External storage with each IP address is a single disk, and each disk is grouped to be recognized as a single storage. At this point, the storage configuration selection technique is used to allow the user to configure the storage to be recognized as a single storage using the storage manager. For example, if four external storages are recognized and there are A, B, C, and D storages, then A-B can be grouped into one storage, C-D to form one storage.

Clustering storage is also provided with the ability to use storage clusters manager 122 to expand storage using disks from additional hosts when virtual hosts and cloud hosts are added. The storage cluster manager 122 clusters the disks of the hosts that are added to the existing clustered storage to the storage, where the capacity also expands. For example, a new host E is added, and E is added to the clustered storage as A-B to form an expanded storage grouped as A-B-E.

When a new host is added to the clustered storage, the redistribution technology that redistributes the data to each storage that constitutes the storage when the storage is expanded, balances the data stored in each storage.

When E is added with data stored in A and B in the state of A-B, the data in A-B is distributed to E and the balance is adjusted.

Each external storage that makes up the clustered storage still has IP. It then uses multipathing technology that uses this IP to recognize clustered storage as the same storage. The CD-clustered storage has both C's IP and D's IP, and the storage pointed to by C and the storage pointed to by D point to the same storage, with the access path in C and the access path in D pointing to the same storage Pass function. The multipath function is a technique in which a path is created to another point regardless of which point is lost due to the presence of two or more mount points. In the present invention, since the mount point 230 is connected to the external storage, the multipath function corresponds to the IP address.

A virtualization method using an internal disk in a server to provide high availability of the virtual machine of the present invention will be described with reference to the following embodiments.

≪ Example 1 >

FIG. 5 shows a flow of a method of virtualizing using the built-in disk in the server.

When the server operates, the external storage providing unit for recognizing the external storage of the internal disk is activated (S110), and the external storage providing unit 121 confirms the type, configuration, and capacity of the internal disk (S120). Next, the server determines which volume the internal disk is to be mounted in (S130) so that the internal disk of the host can be recognized by the server (S140). A mounted volume is recognized as a local volume on the host and can only be used on that host.

The following is an example of how to cluster the server's internal disks in a virtualization-based storage system.

≪ Example 2 >

6 shows a flowchart of a clustering method of a built-in disk in a server.

First, information is registered so that the external storages generated by the external storage providing unit 121 are used for one cloud or virtualization service (S210). Second, in order to expand the clustered storage in the storage cluster manager unit 122, the storage data is redistributed and the storage weights are calculated to balance the data (S220). Then, according to the calculated storage weights, An integrated mount point is provided (S230). The external storage created in step S210 has an IP address and can access the same storage even though the mount point is different by using multipath technology to recognize the same storage using the IP address.

In the present invention, a distributed method and a replicated method are used to store data on a clustered disk. The distributed method has the advantage of being able to use the maximum capacity of the external storage and the speed is very high. In addition, data is replicated and stored on each host in external storage to provide high availability. Figure 7 shows a conceptual diagram of an internal data replication and distribution structure of a clustered disk. The storage replication / distribution manager unit 123 can select whether to replicate data to the built-in disks in each host, to distribute, to split, and to apply a policy in a complex manner.

The internal disks of a cloud or virtualization host are recognized as external storage and can be bundled together in the form of A-B to be used as a single storage. At this point, you can decide how you want to store the data on each host. The first is replication. The replication method is a technique that stores the original in A when the data is stored in the storage enclosed in the form of A-B, and the same copy as the original is stored in B. The replication method is a technique for securely storing data. Even if a failure occurs in the host A, data is not lost because the data of the replicated host B remains. A distributed method is a technique of storing original data in A and one in B. When storing two files, one file is stored in A and the other file is stored in B . If you want to save multiple files at the same time, it will be divided into external storage connected to the network to speed up the storage and reading of files. The partitioning method is a method of dividing the original data into part A and part B and storing and reading the original data when storing the data in the storage enclosed in the form of A-B. It is a way to speed up the data because it is partitioned when a file is written to or read from the storage. The hybrid method is a technique for using a replication method, a distributed method, and a partition method depending on the nature of storage usage when storing data in the form of A-B storage. It supports replication and distribution at the same time, distribution and partitioning at the same time, and replication, distribution and partitioning at the same time.

In the present invention, an external storage service is provided using a built-in disk in a plurality of virtualization hosts. Therefore, a high availability technology capable of providing a storage service safely in the event of a failure in a host is provided. And a failure recovery conceptual diagram.

Each storage host provides a heartbeat function to ensure that the hosts are currently running and are not at fault. Heartbeat periodically checks whether the host that is a member of the storage is running and whether the service is normal. If there is an abnormality in one or more heartbeats of the virtualization hosts that are members, Heartbeat determines that the host has failed. If a failure occurs, identify which host has failed. Then, the host having the same data as the host is set to handle the failure. Storage replication mode must be used in order to use the high availability technology, and it is possible to use a combination of distribution and partitioning in addition to the replication mode. In order to deal with a case where a failed host has failed and is to join again as a member of the external storage, the present invention uses techniques such as recovery and re-membering of the failed host, data replication, and the like.

It will be understood by those skilled in the art that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. It is to be understood, therefore, that the embodiments described above are to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, All changes or modifications that come within the scope of the equivalent concept are to be construed as being included within the scope of the present invention.

100: External Storage System 110: Virtualization / Cloud Solution Division
120: storage high availability manager section 121: external storage provisioning section
122: Storage Cluster Manager Part 123: Storage Replication / Distribution Manager
124: high availability processing manager unit 130: external storage unit
210: external access point unit 220: volume mount unit
230: Mount point 310: Storage weight calculation unit
320: Integrated Mount Point

Claims (12)

In a virtualization-based storage system using an internal disk in a server to provide high availability of a virtual machine,
A virtualization and clutter solution part for connecting external storage by registering information on recognized integrated mount points;
Storage high availability manager responsible for clustering, replication and distribution, and fault handling for high availability of storage;
And external storage to support high availability of storage.
The storage high availability manager
A storage cluster manager unit for clustering external storage of the host; And an external storage providing unit for creating an integrated mount point so that the clustered external storage can be mounted as one unit.
And,
The external storage providing unit
A volume mount section for determining which disk is mounted on which volume so that the internal disk can be recognized by the server; And an external access point unit for determining which external volume the mounted volume is to be recognized and connected to, and an external access point unit for using the built-in disk.
The method according to claim 1,
The storage high availability manager
A storage replication and distribution manager unit for determining whether data is to be copied, distributed, or divided into internal disks in each host;
Further comprising a high availability processing manager unit for providing the storage service unchanged by replicating all the same data to the host having no failure in the occurrence of a failure of the host.
The external storage system of claim 2, wherein the external storage unit is configured to cluster external storage of each host through the storage cluster manager unit and to provide an integrated mount point so that the clustered external storage can be mounted as one unit Virtualization-based storage system with built-in disk.
delete The storage system according to claim 1, further comprising a storage weight calculation unit for balancing data by redistributing storage data to expand the clustered storage in the storage cluster manager unit system.
The method of claim 5, wherein the storage high availability manager unit uses a heartbeat function for determining a failure of the host and sets a host having the same data as the failure host to process the failure. Virtualization-based storage systems.
delete delete delete delete delete delete

Images