CN110502364B - Cross-cloud backup recovery method for big data sandbox cluster under OpenStack platform - Google Patents
Cross-cloud backup recovery method for big data sandbox cluster under OpenStack platform Download PDFInfo
- Publication number
- CN110502364B CN110502364B CN201810472622.2A CN201810472622A CN110502364B CN 110502364 B CN110502364 B CN 110502364B CN 201810472622 A CN201810472622 A CN 201810472622A CN 110502364 B CN110502364 B CN 110502364B
- Authority
- CN
- China
- Prior art keywords
- cluster
- big data
- virtual machine
- backup
- cloud
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/14—Error detection or correction of the data by redundancy in operation
- G06F11/1402—Saving, restoring, recovering or retrying
- G06F11/1446—Point-in-time backing up or restoration of persistent data
- G06F11/1448—Management of the data involved in backup or backup restore
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
- G06F2009/4557—Distribution of virtual machine instances; Migration and load balancing
Abstract
The invention belongs to the technical field of cloud computing and big data, and relates to a cross-cloud backup recovery method for a big data sandbox cluster under an OpenStack platform. The method comprises the following steps: in a big data sandbox cluster backup stage, creating a multi-node virtual machine snapshot and storing the snapshot as a mirror image, and extracting and storing cluster information in an OpenStack-Sahara database; in a cross-cloud transmission stage, reliable transmission of the cluster backup image file is carried out by using an improved UDT protocol; in a big data sandbox cluster recovery stage, the virtual machine backup mirror image is uploaded, the cluster is restarted by using Sahara, and starting of Hadoop and Spark big data service in the cluster is achieved by modifying a cluster plug-in. The method can solve the problems of backup migration and data recovery of the virtual machine and the big data sandbox cluster in order to prevent system or hardware errors of the physical nodes in a multi-cloud big data platform environment.
Description
Technical Field
The invention belongs to the technical field of cloud computing and big data, and particularly relates to a cross-cloud backup recovery method for a big data sandbox cluster under an OpenStack platform.
Background
The prior art discloses that in the cloud computing era, computing, storage and network resources can be used as required, and billed on a per-volume basis, as with traditional energy sources; the virtualization technology aggregates server hardware resources to provide computing resources to users in the form of virtual machines; however, practice shows that unstable factors inside and outside the data center affect the stability and reliability of computing resource supply, and thus, sudden virtual machine failure can cause problems such as interruption of client application, damage of virtual disk data and the like; more seriously, the destructive virtual disk data corruption often results in a virtual machine that cannot be restarted for recovery. Cloud computing platforms typically support backup and restore of a single virtual machine node.
With the development of big data technology and the continuous increase of data volume, a single virtual machine cannot meet the requirement of computing resources, a plurality of virtual machine instances can be arranged into a big data processing cluster by the Sahara component of the OpenStack cloud platform, and a big data computing platform such as Hadoop, spark and the like is operated on the cluster. In the face of big data transaction and processing requests of a plurality of users with different data sizes, different performances and cost requirements, the system needs to adapt to various application characteristics, dynamically and efficiently completes automatic and rapid configuration of a plurality of computing clusters (Hadoop and Spark), forms a special space of each user, namely a big data processing sandbox cluster, and performs safe, reliable and efficient big data analysis and processing in the relatively isolated sandbox; thus, a big data processing sandbox cluster is a self-contained, independent virtual environment dedicated space that carries specific big data processing tasks, relatively securely isolated from other computing resources, based on the big data platform's shared physical and virtual computing/storage resources. The backup recovery function of a single node of an original virtual machine on an OpenStack platform in the prior art cannot meet the cross-cloud disaster recovery requirement under a multi-node scene of a big data sandbox cluster.
Based on the current situation of the prior art, the inventor of the application aims to design a set of cross-cloud sandbox cluster backup and recovery method aiming at a big data virtual machine sandbox cluster.
Disclosure of Invention
The invention aims to provide an overall cross-cloud backup recovery scheme aiming at a big data sandbox cluster under an OpenStack platform based on the current situation of the prior art. In particular to a cross-cloud backup recovery method for a big data sandbox cluster under an OpenStack platform.
The method comprises the following steps: in a big data sandbox cluster backup stage, creating a multi-node virtual machine snapshot and storing the snapshot as a mirror image, and extracting and storing cluster information in an OpenStack-Sahara database; in a cross-cloud transmission stage, reliable transmission of the cluster backup image file is carried out by using an improved UDT protocol; in a big data sandbox cluster recovery stage, uploading a virtual machine backup mirror image, restarting a cluster by using Sahara, and realizing the starting of Hadoop and Spark big data services in the cluster by modifying a cluster plug-in. The method can solve the problems of migration and data recovery of the large data sandbox virtual machine cluster when a physical node has a system or hardware error in an OpenStack cloud platform environment. By using the method, the stability of the big data computing environment can be effectively ensured, and a help and guarantee method is provided for disaster recovery of the big data cluster.
The invention is implemented by the following technical scheme:
the invention provides a cross-cloud backup recovery method of a big data sandbox cluster under an OpenStack platform, which is characterized in that the backup and recovery of the whole cross-cloud platform of the big data sandbox virtual machine cluster are realized through the following steps:
1) Big data sandbox cluster backup phase
After receiving a backup request, creating snapshots of all virtual machines in the big data sandbox cluster, and storing mirror image files; the database backup module extracts cluster information from the database and stores the cluster information in a dictionary form;
2) Cross-cloud transmission big data sandbox cluster mirror image
Cross-cloud data transmission of the big data sandbox cluster backup files is carried out by using an improved UDT protocol;
3) Big data sandbox cluster recovery phase
After receiving a recovery request, uploading a virtual machine backup file into a mirror image, simultaneously recovering information of a node group in a database, after the virtual machine is started, starting a cluster virtual machine instance by using a big data sandbox cluster recovery template plug-in, then starting an original process and service of each node, and completing cluster recovery operation.
In the invention, the cluster backup comprises the following procedures:
1) Storing cluster information, including: cluster configuration, node group, virtual machine information and key pair;
2) Storing cluster template information including node group configuration of a cluster and service operated by each node;
3) Storing configuration information of the registered mirror image, including mirror image basic information and cluster plug-in version;
4) Creating a virtual machine instance snapshot, downloading a snapshot image and converting the format.
In the invention, in the process of creating the virtual machine instance snapshot, after the snapshot is created, qemu is used for converting the mirror image format of the virtual machine from raw to qcow2 so as to improve the mirror image uploading/downloading efficiency.
In the invention, in the improved UDT protocol, the ACK control packet and the loss retransmission list of the UDT protocol are utilized for optimization, so that the reliability of the UDT protocol is improved; aiming at the problem that the UDT protocol takes packet loss as a congestion control signal in a high-speed network, the transmission speed is adjusted by calculating the network bandwidth capacity and the round-trip delay.
In the present invention, the recovery flow in the cluster recovery method is as follows:
1) Confirming whether the number of the mirror images of the virtual machine to be recovered and the type of the virtual machine are matched with the cluster information or not;
2) Uploading a key pair and a virtual machine instance type;
3) Uploading the cluster virtual machine instance mirror image;
4) Starting each node virtual machine instance;
5) Cluster configuration and node group information are imported into a Sahara database;
6) And restarting the original process and service of each node of the cluster by using the Sahara component API.
In the invention, in the cluster plug-in, an OpenStack-Sahara API is called to log in each node virtual machine instance and start the original Hadoop and Spark big data service process.
More specifically, in the present invention,
1. backup big data sandbox cluster
The OpenStack can perform an operation of creating a snapshot on the virtual machine, and is used for storing the current running state of the virtual machine and data in the virtual machine;
the sandbox cluster is composed of a plurality of virtual machine examples, taking a Hadoop cluster as an example, the virtual machines in the cluster are generally divided into a Master node and a plurality of Slave nodes, services such as NameNode and Resourcemanager are operated on the Master node, and services such as DataNode are operated on the Slave nodes; the Sahara component is responsible for scheduling the mutual communication among the nodes in the same cluster, uniformly stores the information of all the nodes in a database, monitors the health state of each node in the cluster and maintains the stable operation of the cluster; therefore, for the backup of the cluster, the method saves the data stored in each virtual machine node on one hand, and saves the overall scheduling information of the cluster in the Sahara database on the other hand; the steps include:
1) After receiving the backup request, according to the ID of the cluster, finding out the corresponding cluster information in the data, and storing the cluster information in a dictionary form, wherein the method comprises the following steps: cluster configuration, node group, virtual machine information and key pair;
2) Saving cluster template information, including node group configuration of the cluster, hadoop operated by each node, and Spark service;
3) Storing configuration information (including image basic information and cluster plug-in version) of the registered image and security group information;
4) Creating snapshots of all virtual machine instances of the cluster by using an OpenStack Nova component, and storing the snapshots of the virtual machines in a Glance component in a mirror image mode by the Nova; downloading a snapshot image file from the Glance component, wherein the size of the image file is the size of a disk of the virtual machine but not the size of actual data, converting the image in the raw format into qcow2 by using a qemu-img tool, and the size of the converted image is the size of the data file in the disk of the virtual machine, so as to complete backup operation;
2. cross-cloud transmission big data sandbox cluster mirror image
Performing backup recovery operation of the cluster among different cloud platforms, and needing to perform a large amount of data transmission; the UDT protocol is a protocol suitable for large data transmission of a high-speed wide area network, has the advantages of high transmission speed, stability and the like in a gigabit network environment, but has the problems of low transmission performance, continuous packet loss and the like in a gigabit network environment;
the improved UDT protocol is obtained through the following improvement of a retransmission mechanism and optimization of a file writing process for the UDT protocol. Aiming at the problem of overtime retransmission caused by the packet loss of a UDT (UDP-based Data Transfer Protocol) Protocol, the improved UDT Protocol is optimized by an ACK control packet and a loss retransmission list in the UDT Protocol based on BDP (Bandwidth-Delay Product, network Bandwidth time-Delay Product), so that the reliability of the UDT Protocol is improved; the steps include:
1) Retransmission mechanism improvements
The method improves the reliability control algorithm of the UDT, so that the UDT can find the lost retransmission data packet in the link in time, and fully utilizes the ACK control packet information to avoid overtime retransmission and retransmission of the received data packet; the receiving end counts the times of the same ACK control packet by setting a counter, retransmits and sends a NAK control packet to inform the packet loss condition of the transmitting end after reaching the current network specified threshold value, and retransmits the unacknowledged data packet, as shown in FIG. 2;
in the present invention,
in Algorithm 1, the input is the sequence number { ACK1, ACK2, \ 8230;, ACKn }, TA } of a series of ACK control packets sent by the receiver same Indicating a threshold for sending the same ACK control packet, RA same And counting the times of the same ACK serial number by the counter, increasing the counter if the currently sent ACK control packet ACKi is the same as the previously sent ACKi-1, and clearing in time if the currently sent ACK control packet ACKi is not the same as the previously sent ACKi-1. When RA is same Over TA same When the data packet is received, reading the data packet in the receiving end loss list RLL, and retransmitting the NAK control packet (preferably selecting the continuous data packet with a small sequence number); defining network bandwidth delay product BDP = Bw btl ×Rtt link Through the network bandwidth Bw btl And round trip delay Rtt link Representing the maximum value of the data quantity which can be transmitted in the link, thereby controlling the size of the congestion window and limiting the data quantity transmitted in the link; TA (TA) same The data packet retransmitted when the BDP is large is related to the BDP value of the current network environmentPossibly also transmitted in the network, TAs ame A larger value should be obtained;
2) File write process optimization
The method comprises the following steps: continuously storing the data packets which reach a fixed number in a continuous sequence into a circular queue, and writing the data packets into a file together; firstly, setting the maximum fixed number of continuously arriving data packets stored each time as FP; then, when receiving FP data packets with continuous sequence numbers, the UDT receiving end stores the FP data packets into a first unit of a circular queue; finally, writing the data information of all the data packets in the unit into the file through the one-time file writing system call in the operation of writing the file;
in the present invention,
in the optimized main thread of the UDT receiving end, data in the receiving buffer area is written into a file in sequence, and the specific writing process is as follows: firstly, acquiring a pointer of the minimum sequence number of unwritten files in a receiving buffer area, and finding a position in a ring queue pointed by the pointer, wherein the mark bit of the position is 0, which represents the starting position of continuous storage; then calculating the size of the continuous storage area of the block, if the mark bit of the subsequent data packet is 1, adding the size of the data packet into the sum, and if the mark bit of the subsequent data packet is 0, indicating that the continuous storage area is interrupted, writing the continuous storage area of the block into a file; finally, updating the positions of the serial numbers which are confirmed to be received in the receiving buffer area, and setting the sum of the sizes of the statistical continuous data packets as 0 byte;
in the present invention, the implementation process of writing the data packet in the data buffer area into the file is shown as algorithm 3:
3. restoring big data sandbox cluster
The cluster cross-cloud backup recovery can cause the original network environment to change, and the OpenStack can allocate new instance IDs, IP addresses and the like to newly-established virtual machine instances, so that the completeness of all node data is ensured, and meanwhile, after the recovery operation is finished, normal communication among all nodes of the cluster can be ensured, and the original service can run normally; the steps comprise:
1) After receiving a recovery request, determining whether the number of the mirror images of the virtual machine to be recovered, the type of the virtual machine and the cluster information are matched or not;
2) When the backup recovery of a cross-cloud platform is involved, the authentication information of a user can be changed, the user can log in by using an original key pair of a cluster, and a new replacement original key pair can be created;
3) Uploading the cluster virtual machine instance mirror image;
4) Starting a virtual machine instance, and replacing original cluster information in a database by using information such as a new instance ID, an IP address and the like of a node;
5) Configuring a cluster, and importing node group information into a Sahara database;
6) And starting the corresponding Hadoop and Spark service processes on each node by using a Sahara component API, and finishing the recovery operation.
The method disclosed by the invention is verified by practice, and the result shows that the problems of migration and data recovery of the large data sandbox virtual machine cluster when a physical node has a system or hardware error in the OpenStack cloud platform environment can be solved. By using the method, the stability of the big data computing environment can be effectively ensured, and a help and guarantee method is provided for disaster recovery of the big data cluster.
Drawings
Fig. 1 is an overall work flow diagram.
Fig. 2 is a flow chart of the receiving end retransmitting NAK control packet.
Detailed Description
Example 1
In the experimental part of the invention, a large data sandbox cluster cross-cloud backup recovery operation is carried out on two OpenStack cloud platforms once.
The cloud platform environment is as follows:
1. cluster backup OpenStack platform: the OpenStack Ocata version, one control node and five computing nodes. The ip address of the control node is 10.141.105.19.
2. And cluster recovery OpenStack platform: the OpenStack Queens edition, one control node and five computing nodes. The ip address of the control node is 10.141.105.14.
The cluster information to be operated is as follows:
master node, 1:
virtual machine configuration: a CPU:2, memory: 4GB, 40GB of disk
The running service process:
·NameNode
·SecondaryNameNode
·ResourceManager
·HistoryServer
·oozie
·HiveServer
·Spark History Server
slave node, 3:
virtual machine configuration: a CPU:2, memory: 4GB, 40GB of disk
The running service process:
·DataNode
·NodeManager
first a request containing the cluster ID is sent to http://10.141.105.19/sandbox/backup to start the backup operation.
In the process, firstly, according to a cluster ID, cluster information including a cluster template, a cluster node group template, a cluster key pair, a virtual machine type of a cluster node, cluster security group information and cluster mirror image information is searched and stored from a database, and the information is stored in a file in a json format for re-acquisition during recovery;
then, starting to backup a virtual machine instance of the cluster node, calling an API (application programming interface) of an OpenStack Nova component according to a virtual machine ID (identity) corresponding to the cluster node, creating a snapshot mirror image of the virtual machine, and acquiring a mirror image file with a raw format through the API of the OpenStack Glance component, wherein the size of the raw format mirror image is equal to the preset size of a disk of the virtual machine, and in the example, the size of the disk of the virtual machine is 40GB;
next, using a qemu-img tool to convert the raw format mirror image into a qcow2 format mirror image, after conversion, changing the size of the mirror image file into the actual size of the data file in the disk, then using a UDT protocol to transmit the cluster information file and the virtual machine mirror image file to a host with an IP address of 10.141.105.14, and preparing for recovery operation;
after the file transmission is finished, sending a recovery request comprising a cluster ID, a Neutron network ID and key pair information to http://10.141.105.14/sandbox/recovery, starting a cluster recovery phase,
uploading a mirror image of a virtual machine to Glance, recovering information of a key pair, a node group and a cluster in a database once, updating an IP address of each node according to Nova scheduling, and calling a Sahara API to restart an original service process of the cluster to complete recovery operation;
the result shows that the method can solve the problems of migration and data recovery of the large data sandbox virtual machine cluster when a physical node has a system or hardware error in the OpenStack cloud platform environment.
Claims (6)
1. A cross-cloud backup recovery method of a big data sandbox cluster under an OpenStack platform is characterized in that the backup and recovery of the big data sandbox cluster are realized through a cross-cloud data center through the following steps:
1) Big data sandbox cluster backup phase
After receiving a backup request, creating snapshots of all virtual machines in the big data sandbox cluster, and storing mirror image files; the database backup module extracts cluster information from the database and stores the cluster information in a dictionary form;
2) Cross-cloud transmission big data sandbox cluster mirror image
Cross-cloud data transmission of the large data sandbox cluster backup files is carried out by using an improved UDT protocol;
3) Big data sandbox cluster recovery phase
After receiving a recovery request, uploading a virtual machine backup file into a mirror image, simultaneously recovering information of a node group in a database, after the virtual machine is started, starting a cluster virtual machine instance by using a big data sandbox cluster recovery template plug-in, then starting an original process and service of each node, and completing cluster recovery operation.
2. The cross-cloud backup recovery method for the big data sandbox cluster under the OpenStack platform according to claim 1, wherein the cluster backup comprises the following procedures:
1) Storing cluster information, including: cluster configuration, node group, virtual machine information and key pair;
2) Saving cluster template information including the configuration of a cluster node group and the service operated by each node;
3) Storing configuration information of the registered mirror image, including mirror image basic information and cluster plug-in version;
4) Creating a virtual machine instance snapshot, downloading a snapshot image and converting the format.
3. The cross-cloud backup recovery method for the big data sandbox cluster under the OpenStack platform according to claim 2, wherein in creating the snapshot of the virtual machine instance, after the snapshot is created, qemu is used to convert the image format of the virtual machine from raw to qcow2 so as to improve the image uploading/downloading efficiency.
4. The cross-cloud backup recovery method for the big data sandbox cluster under the OpenStack platform according to claim 1, wherein in the improved UDT protocol, the reliability of the UDT protocol is improved by optimizing the ACK control packet and the lost retransmission list of the UDT protocol; aiming at the problem that the UDT protocol takes packet loss as a congestion control signal in a high-speed network, the transmission speed is adjusted by calculating the network bandwidth capacity and the round-trip delay.
5. The cross-cloud backup recovery method for the big data sandbox cluster under the OpenStack platform according to claim 1, wherein the cluster recovery comprises the following procedures:
1) Confirming whether the number of the mirror images of the virtual machines to be recovered is matched with the type of the virtual machine and the cluster information;
2) Uploading a key pair and a virtual machine instance type;
3) Uploading the cluster virtual machine instance mirror image;
4) Starting each node virtual machine instance;
5) Cluster configuration and node group information are imported into a Sahara database;
6) And restarting the original process and service of each node of the cluster by using the Sahara component API.
6. The cross-cloud backup recovery method for the big data sandbox cluster under the OpenStack platform according to claim 2, wherein the cluster plugin calls an OpenStack-Sahara API to log in each node virtual machine instance and start an original Hadoop and Spark big data service process.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810472622.2A CN110502364B (en) | 2018-05-17 | 2018-05-17 | Cross-cloud backup recovery method for big data sandbox cluster under OpenStack platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810472622.2A CN110502364B (en) | 2018-05-17 | 2018-05-17 | Cross-cloud backup recovery method for big data sandbox cluster under OpenStack platform |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110502364A CN110502364A (en) | 2019-11-26 |
| CN110502364B true CN110502364B (en) | 2023-03-14 |
Family
ID=68583955
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810472622.2A Active CN110502364B (en) | 2018-05-17 | 2018-05-17 | Cross-cloud backup recovery method for big data sandbox cluster under OpenStack platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110502364B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111078667B (en) * | 2019-12-12 | 2023-03-10 | 腾讯科技（深圳）有限公司 | Data migration method and related device |
| CN111726403B (en) * | 2020-06-11 | 2021-01-29 | 和宇健康科技股份有限公司 | Cross-cloud-platform big data management method and system |
| CN111797059A (en) * | 2020-07-07 | 2020-10-20 | 优刻得科技股份有限公司 | Method for online cloud host migration across clouds, migration controller and cloud server |
| CN112084491A (en) * | 2020-08-26 | 2020-12-15 | 天津七一二通信广播股份有限公司 | Sandbox mechanism-based cluster virtual user system implementation method |
| CN112929609A (en) * | 2021-02-05 | 2021-06-08 | 中国电信集团工会上海市委员会 | Video monitoring system based on cloud platform |
| CN116049136B (en) * | 2022-12-21 | 2023-07-28 | 广东天耘科技有限公司 | Cloud computing platform-based MySQL cluster deployment method and system |
| CN116028277B (en) * | 2023-03-27 | 2023-07-28 | 广州智算信息技术有限公司 | Database backup method and system based on CDC mode |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102521071A (en) * | 2011-11-24 | 2012-06-27 | 广州杰赛科技股份有限公司 | Private cloud-based virtual machine maintaining method |
| CN102834822A (en) * | 2010-04-12 | 2012-12-19 | 微软公司 | Express-full backup of a cluster shared virtual machine |
| CN106557355A (en) * | 2016-12-01 | 2017-04-05 | 北京奇虎科技有限公司 | The generation method and generating means of virtual machine image |
| US9639428B1 (en) * | 2014-03-28 | 2017-05-02 | EMC IP Holding Company LLC | Optimized backup of clusters with multiple proxy servers |
| CN107197022A (en) * | 2017-06-02 | 2017-09-22 | 华南理工大学 | OpenStack storage optimization methods and system |
| US9842124B1 (en) * | 2013-03-13 | 2017-12-12 | EMC IP Holding Company LLC | Highly available cluster agent for backup and restore operations |
-
2018
- 2018-05-17 CN CN201810472622.2A patent/CN110502364B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102834822A (en) * | 2010-04-12 | 2012-12-19 | 微软公司 | Express-full backup of a cluster shared virtual machine |
| CN102521071A (en) * | 2011-11-24 | 2012-06-27 | 广州杰赛科技股份有限公司 | Private cloud-based virtual machine maintaining method |
| US9842124B1 (en) * | 2013-03-13 | 2017-12-12 | EMC IP Holding Company LLC | Highly available cluster agent for backup and restore operations |
| US9639428B1 (en) * | 2014-03-28 | 2017-05-02 | EMC IP Holding Company LLC | Optimized backup of clusters with multiple proxy servers |
| CN106557355A (en) * | 2016-12-01 | 2017-04-05 | 北京奇虎科技有限公司 | The generation method and generating means of virtual machine image |
| CN107197022A (en) * | 2017-06-02 | 2017-09-22 | 华南理工大学 | OpenStack storage optimization methods and system |
Non-Patent Citations (1)
| Title |
|---|
| Construction and application of Red5 cluster based on OpenStack;Jiaqing Wang et al.;《AIP Conference Proceedings》;20170803;第1-8页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110502364A (en) | 2019-11-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110502364B (en) | Cross-cloud backup recovery method for big data sandbox cluster under OpenStack platform | |
| US10645019B2 (en) | Relaxed reliable datagram | |
| JP6490310B2 (en) | Networking technology | |
| CN110297801B (en) | System and method for just-in-one transaction semantics of transaction system based on fault-tolerant FPGA | |
| WO2021121370A1 (en) | Message loss detection method and apparatus for message queue | |
| US9424325B2 (en) | Recording medium, distribution controlling method, and information processing device | |
| US8180928B2 (en) | Method and system for supporting read operations with CRC for iSCSI and iSCSI chimney | |
| US20050281287A1 (en) | Control method of communication system, communication controller and program | |
| US11025564B2 (en) | RDMA transport with hardware integration and out of order placement | |
| WO2021142971A1 (en) | Transmission rate control method and apparatus, computer system, and readable storage medium | |
| US10491329B1 (en) | Transfer of data-redundancy encoded data via unreliable, connectionless protocol | |
| US11303737B2 (en) | Method and device for data transmission | |
| WO2017088572A1 (en) | Data processing method, device, and system | |
| US20200272579A1 (en) | Rdma transport with hardware integration | |
| US10521317B1 (en) | Compressing data to be replicated utilizing a compression method selected based on network behavior | |
| CN113986501A (en) | Real-time database API (application program interface) uninterrupted calling method, system, storage medium and server | |
| US20040267960A1 (en) | Force master capability during multicast transfers | |
| WO2016101409A1 (en) | Data switching method, device and system | |
| US11444882B2 (en) | Methods for dynamically controlling transmission control protocol push functionality and devices thereof | |
| US10735515B2 (en) | Layered distributed storage system and techniques for edge computing systems | |
| US20220217093A1 (en) | Sequence Number Synchronization Method and Apparatus | |
| CN115065694A (en) | Cloud storage data transfer uploading system, method, equipment and medium | |
| CN114885007A (en) | Method and electronic device for real-time strong consistency session synchronization | |
| US11336592B2 (en) | Flexible link level retry for shared memory switches | |
| JP5081798B2 (en) | Data transfer method, path configuration management apparatus, and path configuration management program |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |