CN105262820A - Cluster multi-computer mutual backup method based on Linux operating system - Google Patents

Abstract

The invention discloses a method for mutual backup of multiple machines in a cluster based on a Linux operating system, which belongs to a method for mutual backup of clusters. The problem of combining with CTDB in the mutual backup of multiple machines in the cluster is adopted. The technical solution is: in the Linux operating system, use the shared storage block of GlusterFS and cooperate with the service monitoring and switching service of CTDB to realize the mutual backup of multiple machines in the cluster and provide external Floating IP, FTP service, NFS service, Samba service and application service; including the following deployment steps: (1), configure the hosts file: write the information of all client nodes into the hosts file, and ensure that the nodes can realize mutual Password access; (2), configure GlusterFS: use the GlusterFS cluster file system to provide shared storage space for all nodes, and provide shared storage space for lock and status for CTDB, and create CTDB volumes and NAS volumes; (3), configure CTDB.

Description

A method of cluster multi-computer mutual backup based on Linux operating system

technical field

The invention relates to a method for mutual backup of clusters, in particular to a method for mutual backup of clusters based on a Linux operating system.

Background technique

With the rapid development of cloud computing, the Internet, and big data, people have higher and higher requirements for machine performance, especially the high availability of clusters. A cluster is a computer group that is connected by a series of independent computers through high-speed Internet and provides external services through unified management. Cluster high availability usually describes a system that has been specially designed to minimize downtime to ensure cluster business continuity.

Generally, there are three working modes for cluster high availability: master-slave mode, dual-machine duplex mode, and cluster working mode;

(1) Master-slave mode: the main machine works, and the standby machine monitors. When the main machine fails, the service switches to the standby machine; when the main machine recovers, the service switches back to the main machine;

(2) Two-machine duplex mode: Two hosts run their respective services at the same time and monitor each other, and can ensure that when one fails, the other can take over;

(3) Cluster working mode: Multiple hosts in the cluster work together, each running one or more services. After any host fails, its business can be taken over by other hosts.

The Linux operating system is a set of Unix-like operating systems that are free to use and spread freely. It is a multi-user, multi-tasking, multi-threading and multi-CPU operating system based on POSIX and UNIX. It can run major UNIX utility software, application programs and network protocols. It supports 32-bit and 64-bit hardware. Linux inherits Unix's network-centric design idea and is a multi-user network operating system with stable performance.

GlusterFS (GlusterFileSystem) is a free software. As an open source distributed file system, it is applied to cluster systems and has good scalability. It can support PB storage capacity through expansion, and physically distributes the data with the help of TCP/IP or InfinibandRDMA network. It mainly solves the application of cluster file system. The main functions include: network storage, joint storage (integration of storage space on multiple nodes), redundant backup, load balancing of large files, etc.

The full name of CTDB is sambaCTDB, which is implemented based on the TDB database used by samba. The database is used to store temporary files including samba or other applications. CTDB provides a TDB type database for multiple physical hosts in the cluster. The database provides continuous data information for the entire cluster to ensure the business continuity of the cluster, and the database is automatically maintained by CTDB. In addition, in order to achieve high availability, CTDB provides functions such as node monitoring, failover, and IP takeover. It is worth mentioning that CTDB supports switching of system or application management scripts during high availability.

Samba is a free software that implements the SMB protocol on Linux and UNIX systems, and consists of server and client programs. SMB (ServerMessagesBlock, Information Service Block) is a communication protocol for sharing files and printers on a LAN. It provides sharing services for resources such as files and printers between different computers in the LAN. The SMB protocol is a client/server protocol through which the client can access the shared file system, printers and other resources on the server. By setting "NetBIOS over TCP/IP", Samba can not only share resources with LAN hosts, but also share resources with computers all over the world.

For the application requirements of high-performance clusters, big data clusters, and massive storage clusters, how to deploy multiple servers under the Linux operating system, how to achieve multi-machine mutual backup and high availability of clusters with uninterrupted business, and how to combine GlusterFS and CTDB In cluster multi-machine mutual backup, it is an urgent problem to be solved at present.

Contents of the invention

The technical task of the present invention is to provide a method for multi-machine mutual backup of clusters based on the Linux operating system, to solve how to realize multi-machine mutual backup and high availability of clusters with uninterrupted business, and how to combine GlusterFS and CTDB in cluster multi-machine Problems in machine mutual backup.

Technical task of the present invention is realized in the following manner,

A method for mutual backup of multiple machines in a cluster based on the Linux operating system. In the Linux operating system, the shared storage block of GlusterFS is used together with the service monitoring and switching service of CTDB to realize the mutual backup of multiple machines in the cluster, achieve high availability of the cluster, and provide Floating IP, FTP service, NFS service, Samba service, and application service; including the following deployment steps:

(1) Configure the hosts file: write the information of all client nodes into the hosts file, and ensure that the nodes can achieve password-free access to each other;

(2), configure GlusterFS: use the GlusterFS cluster file system to provide shared storage space for all nodes, and provide lock and status shared storage space for CTDB, and create CTDB volumes and NAS volumes;

(3) To configure CTDB, the configuration steps are as follows:

(a), configure the ctdb file, on the client1 node, modify the ctdb file in the /mnt/gluster directory;

(b), configure the nodes file, on the client1 node, modify the cluster node file (nodes) under the /mnt/gluster directory;

(c), modify the public_addresses file, define the floating IP, and modify the public network ip file (public_addresses) on the client1 node;

(d), modify the nfs file, on the client1 node, modify the file nfs under the /mnt/gluster directory;

(e) Change the export file;

(f), modify directory permissions, on all client nodes, modify /mnt/icfs directory permissions;

(g) Clean up the original configuration and delete the original configuration file information of nfs and ctdb on all client nodes;

(h), establish a connection, and establish soft connections to these files on all client nodes, so that these configuration files can be loaded when the software starts;

(i) Start ctdb, and start the ctdb process on each node of the interface node cluster.

The steps to create CTDB volume and NAS volume in step (2) are as follows:

(a) Start the service;

(b) Create a storage pool;

(c), create a volume, the steps are as follows: First, create a volume test-volume, test-volume is used to store the lock file of ctdb, smb, nfs and ctdb configuration file information; then, automatically create /root on the two nodes /gluster directory; then, replica ensures that when one node writes data, it can be synchronized to another node;

(d), mount the unified storage volume, the steps are as follows: first, create the mount directory on the two interface nodes; then, perform the mount operation on the two client nodes (apply their own IP); finally, mount Mount the volume of the machine, after mounting, you can use the df command to check whether the mount is successful;

(e), save the configuration file: save the configuration file required by the interface node cluster on the unified storage volume, including the lock file (lockfile), samba configuration file (smb.conf), nfs configuration file (exports), ctdb configuration file (ctdb ), cluster node files (nodes), public network IP files (public_addresses).

Floating IP: as the business access entrance of all clusters,

FTP service: provide file upload and download services,

NFS service: Provides file sharing services between Linux operating systems,

Samba service: Provide Linux operating system to provide file sharing service for Windows operating system,

Application Service: Provides services that require configuration documents.

A kind of method for the mutual backup of cluster multi-machines based on Linux operating system of the present invention has the following advantages:

1. Through the deployment of multiple servers under the Linux operating system, the current mainstream Linux operating system has dual-machine master-backup, multi-machine master-backup, and uninterrupted business functions;

2. Using the idea of shared storage blocks of GlusterFS, combined with CTDB's service monitoring and switching services, realizes high-availability solutions, and provides external floating IP, FTP services, NFS services, Samba services and application services;

3. A high-availability cluster is realized through the deployment method of the present invention, which can make the overall service of the cluster as available as possible, thereby reducing the loss caused by the error-proneness of computer hardware and software. If a node fails, its backup node will take over its duties within seconds. Therefore, for the user, the cluster never goes down. The main function of high-availability cluster software is to realize the automation of fault detection and business switching. A high-availability cluster with only two nodes is also called dual-system hot standby, that is, two servers are used to back up each other. When one server fails, another server can assume the service task, so that the system can automatically ensure that the system can continue to provide external services without manual intervention. Dual-machine hot backup is only one type of high-availability cluster. The high-availability cluster system can support more than two nodes, provide more and more advanced functions than dual-machine hot backup, and better meet the changing needs of users;

4. Through the deployment method of the present invention, dual-machine active-standby and multi-machine clusters are realized, and high-availability clusters are realized; the present invention does not involve data security, and data security is supported by the underlying storage, which only provides continuity of services .

Description of drawings

The present invention will be further described below in conjunction with the accompanying drawings.

Accompanying drawing 1 is the structure diagram of GlusterFS+CTDB of a kind of cluster multi-computer backup method based on Linux operating system.

detailed description

Referring to the accompanying drawings and specific embodiments, a Linux operating system-based cluster multi-machine mutual backup method of the present invention will be described in detail below.

Example:

As shown in Figure 1, in the figure, the InfiniBand architecture is a "conversion cable" technology that supports multiple concurrent links. In this technology, each link can reach an operating speed of 2.5Gbps. This architecture has a speed of 500MB/s for one link, 2GB/s for four links, and 6GB/s for 12 links. The full name of RDMA (Remote Direct Memory Access) technology is remote direct data access, which was created to solve the delay of server-side data processing in network transmission; the advantage of RDMA is that it can use traditional network hardware to establish the Internet with TCP/IP and Ethernet standards. InfiniBand networks and networks implementing virtual interface architectures support RDMA. Privatenet is an internal network.

A method for mutual backup of clusters based on the Linux operating system of the present invention, in the Linux operating system, utilizes the shared storage block of GlusterFS and matches the service monitoring switching service of CTDB to realize the mutual backup of clusters and achieve high availability of the cluster , provide floating IP, FTP service, NFS service, Samba service and application service; floating IP: as the business access entrance of all clusters, FTP service: provide file upload and download service, NFS service: provide file sharing service between Linux operating systems , Samba service: provide Linux operating system to provide file sharing service for Windows operating system, application service: provide services that require configuration documents; including the following deployment steps:

(1) Configure the hosts file: write the information of all client nodes into the hosts file, and ensure that the nodes can achieve password-free access to each other;

(2) Configure GlusterFS: use the GlusterFS cluster file system to provide shared storage space for all nodes, and provide lock and status shared storage space for CTDB, and create CTDB volumes and NAS volumes. Both CTDB volumes and NAS volumes use stripe volume, select distribute; the steps to create CTDB volumes and NAS volumes are as follows:

(a), start the service; the code operation is as follows:

[rootclient1~]#serviceglusterdstart

[rootclient2~]#serviceglusterdstart

[rootclient1~]#chkconfigglusterdon;

(b) Create a storage pool; the code operation is as follows:

[rootclient1~]#glusterpeerprobeclient2//Add a server to the storage pool

Probesuccessful

[rootclient1~]#glusterpeerstatus//verify membership status

Number of Peers: 1

Hostname: client2

Uuid: 8d9c00aa-6683-4b62-83e7-a4011992add8

State: PeerinCluster(Connected);

(c), create a volume, the steps are as follows: First, create a volume test-volume, test-volume is used to store the lock file of ctdb, smb, nfs and ctdb configuration file information; then, automatically create /root on the two nodes /gluster directory; then, replica ensures that when one node writes data, it can be synchronized to another node; the code operation is as follows:

[rootclient1~]#glustervolumecreatetest-volumereplica2transport

tcpclient1:/root/glusterclient2:/root/gluster

[rootclient1~]#glustervolumestarttest-volume;

(d) To mount a unified storage volume, the steps are as follows: First, create mounting directories on the two interface nodes respectively, and the code operation is as follows:

[rootclient1~]#mkdir/mnt/gluster

[rootclient2~]#mkdir/mnt/gluster;

Then, perform the mount operation on the two client nodes respectively (applying its own IP);

Finally, mount the volume of the machine (you cannot directly write to the /root/gluster directory), the code operation is as follows:

[rootclient1~]#mount-tglusterfs192.168.2.21:/test-volume/mnt/gluster

[rootclient2~]#mount-tglusterfs192.168.2.22:/test-volume/mnt/gluster;

After mounting, you can use the df command to check whether the mounting is successful; (why not add it to /etc/fstab? Because it takes some time to start the glusterfs service when the server starts, adding it to /etc/fstab will not successfully mount loaded);

(e), save the configuration file: save the configuration file required by the interface node cluster on the unified storage volume, including the lock file (lockfile), samba configuration file (smb.conf), nfs configuration file (exports), ctdb configuration file (ctdb ), cluster node files (nodes), public network IP files (public_addresses).

(3) To configure CTDB, the configuration steps are as follows:

(a) Configure the ctdb file. On the client1 node, modify the ctdb file in the /mnt/gluster directory. The code operation is as follows:

CTDB_RECOVERY_LOCK=/mnt/gluster/lockfile

CTDB_PUBLIC_INTERFACE=eth0 (corresponding to the network card connected to the customer network)

CTDB_PUBLIC_ADDRESSES=/etc/ctdb/public_addresses

CTDB_MANAGES_SAMBA=yes

CTDB_MANAGES_WINBIND=yes

CTDB_MANAGES_NFS=yes

CTDB_MANAGES_VSFTPD=yes

CTDB_NODES=/etc/ctdb/nodes

CTDB_DEBUGLEVEL=ERR;

(b) Configure the nodes file. On the client1 node, modify the cluster node file (nodes) in the /mnt/gluster directory. The operation is as follows:

192.168.2.101

192.168.2.102,

These two IPs are the internal network IPs that constitute the interface node cluster;

(c), modify the public_addresses file, define the floating IP, on the client1 node, modify the public network ip file (public_addresses), as follows: 192.168.100.101/24eth0;

(d) Modify the nfs file. On the client1 node, modify the file nfs under the /mnt/gluster directory. The nfs file is modified as follows:

CTDB_MANAGES_NFS=yes

NFS_TICKLE_SHARED_DIRECTORY=/mnt/gluster/nfs-tickles

STATD_PORT=595

STATD_OUTGOING_PORT=596

MOUNTD_PORT=597

RQUOTAD_PORT=598

LOCKD_UDPPORT=599

LOCKD_TCPPORT=599

STATD_SHARED_DIRECTORY=/mnt/gluster/nfs-state

NFS_HOSTNAME="ctdb"

STATD_HOSTNAME="$NFS_HOSTNAME -P

"$STATD_SHARED_DIRECTORY/$PUBLIC_IP" -H /etc/ctdb/statd-callout-p97";

(e), change the export file, the code operation is as follows:

/home*(rw,fsid=09b4f041-8b20-4377-9af3-2658ae2ecba2);

(f), modify directory permissions, modify /mnt/icfs directory permissions on all client nodes, the code is as follows: [rootclient1~]#chmod-R777/mnt/icfs/;

(g) Clean up the original configuration, delete the original configuration file information of nfs and ctdb on all client nodes, and perform the following operations on all client nodes:

[rootclient1~]#rm-rf/etc/samba/smb.conf

[rootclient1~]#rm-rf /etc/sysconfig/ctdb

[rootclient1~]#rm-rf /etc/sysconfig/nfs

[rootclient1~]#rm-rf /etc/exports;

(h), establish a connection, and establish soft connections to these files on all client nodes, so that these configuration files can be loaded when the software starts;

(i), start ctdb, start the ctdb process on each node of the interface node cluster, and perform the following operations on all interface nodes:

[rootclient1~]#servicerpcbindrestart

[rootclient1~]#servicectdbstart

[rootclient2~]#servicerpcbindrestart

[rootclient2~]#servicectdbstart.

Through the above specific implementation manners, those skilled in the technical field can easily realize the present invention. However, it should be understood that the present invention is not limited to the specific embodiments described above. On the basis of the disclosed embodiments, those skilled in the art can arbitrarily combine different technical features to achieve different technical solutions.

Except for the technical features described in the instructions, all are known technologies by those skilled in the art.

Claims (3)

1. A method for mutual backup of multiple machines in a cluster based on the Linux operating system, characterized in that in the Linux operating system, the shared storage block of GlusterFS is used, and the service monitoring and switching service of CTDB is matched to realize the mutual backup of multiple machines in the cluster, and externally provide Floating IP, FTP service, NFS service, Samba service, and application service; including the following deployment steps: (1) Configure the hosts file: write the information of all client nodes into the hosts file, and ensure that the nodes can achieve password-free access to each other; (2), configure GlusterFS: use the GlusterFS cluster file system to provide shared storage space for all nodes, and provide lock and status shared storage space for CTDB, and create CTDB volumes and NAS volumes; (3) To configure CTDB, the configuration steps are as follows: (a), configure the ctdb file, on the client1 node, modify the ctdb file in the /mnt/gluster directory; (b), configure the nodes file, on the client1 node, modify the cluster node file in the /mnt/gluster directory; (c), modify the public_addresses file, define the floating IP, and modify the public network ip file on the client1 node; (d), modify the nfs file, on the client1 node, modify the file nfs under the /mnt/gluster directory; (e) Change the export file; (f), modify directory permissions, on all client nodes, modify /mnt/icfs directory permissions; (g) Clean up the original configuration and delete the original configuration file information of nfs and ctdb on all client nodes; (h), establish a connection, and establish soft connections to these files on all client nodes, so that these configuration files can be loaded when the software starts; (i) Start ctdb, and start the ctdb process on each node of the interface node cluster. 2. A method for mutual backup of clusters based on Linux operating system according to claim 1, characterized in that the steps of creating CTDB volumes and NAS volumes in step (2) are as follows: (a) Start the service; (b) Create a storage pool; (c), create a volume, the steps are as follows: first, create a volume test-volume, test-volume is used to store the lock file of ctdb, smb, nfs and ctdb configuration file information; then, automatically create /root on the two nodes /gluster directory; then, replica ensures that when one node writes data, it can be synchronized to another node; (d), mount the unified storage volume, the steps are as follows: first, create the mount directory on the two interface nodes; then, perform the mount operation on the two client nodes respectively, and apply its own IP; finally, Mount the volume of the local machine, and after mounting, you can use the df command to check whether the mount is successful; (e) Save configuration files: Save the configuration files required by the interface node cluster on the unified storage volume, including lock files, samba configuration files, nfs configuration files, ctdb configuration files, cluster node files, and public network IP files. 3. the method for a kind of cluster multi-machine mutual backup based on Linux operating system according to claim 1, is characterized in that floating IP: as the service access entrance of all clusters, FTP service: provide file upload and download services, NFS service: Provides file sharing services between Linux operating systems, Samba service: Provide Linux operating system to provide file sharing service for Windows operating system, Application Service: Provides services that require configuration documents.