Comparison and analysis of three mainstream server cluster software

Simply put, a cluster (cluster) is a group of computers that provide users with a set of network resources as a whole. These individual computer systems are the nodes (node) of the cluster. An ideal cluster is that the user never realizes the nodes at the bottom of the cluster system, in their view, the cluster is a system, not multiple computer systems. And the administrators of the cluster system can add and revise the nodes of the cluster system at will.

Below on the server commonly used three cluster software to do a comparative analysis of the introduction:

1. ROSE HA server cluster software

In a two-machine hot standby architecture, in addition to considering the switching time, according to the operating environment of each system, including the network system is a single or dual network, the database installation and operation content and user-side equipment is through the WAN Road, regional network access to different users have different needs, and require different switching mode, So choose a different switching mode, can make the user end of the change to achieve the least degree.

ROSE ha based on different industries and different needs of various industries to design a variety of backup models to flexibly adjust the user's best combination and choice.

ROSE ha system mode of operation

In the normal operation of the situation, the host through the redundant detection lines to detect each other, when any of the host error generation, ROSE ha provides rigorous judgment and analysis, to confirm the host error, only to fully start the backup takeover action.

※ Support various operating system platform
※ supports a wide range of Unix platforms (e.g. IBM, DEC, HP, NCR, SUN, SGI, NEC, Siemens, etc.)
※ Support a large number of PC platform Unix systems (such as: Sco/unix, Solraris X86, etc.)
※ Support Various databases: MS, Oracle, Informix, Sysbase, Excheng|, Lotus/nose, DB2, etc.

The takeover action includes

※ FileSystem (File system)
※ Database
※ Network addresses (IP address)
※ Application (AP)
※ System Environment (OS)
※ Fault-tolerant backup operation process

Automatic detection (Auto-detect) phase, by the software on the host through a redundant detection line, through a complex listener. Logical judgment to detect each other's operation, the items examined are:

※ Mainframe hardware (CPU and peripherals)
※ Host Network
※ Host operating system
※ Database engine and other applications
※ Host and disk array connection

In order to ensure the correctness of the detection, and to prevent erroneous judgments, can set the security detection time, including detection time interval, detection times to adjust the safety factor, and by the host's redundant communication lines, the collected messages recorded for maintenance reference.

Automatic switching (auto-switch) phase a host if the other fault, the normal host in addition to continue to carry out the original task, will also be based on a variety of fault-tolerant standby mode to take over the pre-set procedures, and follow-up procedures and services.

The automatic recovery (auto-recovery) phase can be repaired offline after the normal host is replaced by the failed host. After the failure of the host repair, through the redundant communication line and the original normal host connection, automatically switch back to the repair completed host. The whole process is completed by Edi-ha automatically, can also be based on a predetermined configuration, select the response action for semi-automatic or not reply.

Multiple Fault Tolerant backup modes

1, two-machine mutual standby mode
2, master-slave mode of backup
3,hot Standby mode
4, dual network card, single segment type
5, Dual Network mode
6, multi-host standby system

User customization Mode

In a two-machine hot standby architecture, in addition to considering the switching time, according to the operating environment of each system, including the network system is a single or dual network, the database installation and operation content and user-side equipment is through the WAN Road, regional network access to different users have different needs, and require different switching mode, So choose a different switching mode, can make the user end of the change to achieve the least degree.

ROSE ha based on different industries and different needs of various industries to design a variety of backup models to flexibly adjust the user's best combination and choice.

※ Two hosts are connected by a TCP/IP network line and a RS-232 cable line
※ The two hosts are connected with a RAID disk array by a SCSI cable line
※ Each of the two hosts run different operations, independent and mutual assistance
※ Host a failure, Host B automatically take over host a operation
※ Host A's job will be run automatically on Host B
※ Host A's client (client) is to log on Host B again
※ Host a repair, Host B will automatically return a job to host a
※ Customers who are already connected to a on Host B need to log on host a again
※ Host B Fault, host a takes over the operation and data of Host B
※ When the Host B is repaired, host A will return the original operation and data to Host B
※ Two hosts are connected by a TCP/IP network line and a RS-232 cable line
※ Dual hosts are connected by a SCSI cable and raid
※ Host A is master, Host B is slave
※ Host a processing operations and data, Host B as a hot standby machine
※ Host a failure, Host B automatically take over the operation and data of host a
※ Host B takes over the host name of A and the network address (IP)
※ Host A's job will be run automatically on Host B
※ Host a client (client) can continue to run without having to log in again
※ Host B is master, host a repaired as slave, as a hot backup machine
※ 2 hosts recommend using the same specification

Mode 3---Dual-machine hot Standby

※ Two hosts are connected by a TCP/IP network line and a RS-232 cable line
※ Dual hosts are connected by a SCSI cable and raid
※ Host A is master, Host B is slave
※ Host a processing operations and data, Host B as a hot standby machine
※ Host a failure, Host B automatically take over the operation and data of host a
※ Host B takes over the host name of A and the network address (IP)
※ Host A's job will be run automatically on Host B
※ Host a client (client) can continue to run without having to log in again
※ Host a repair, automatically take over the original work and data, Host B continue to make backup machine

Mode 4---Dual-computer dual network adapter (Dual Ethernet Adapter)

※ Dual hosts are connected by a SCSI cable and raid
※ The two hosts run different jobs
※ Each host defines the first and second network ports
※ The first network port in the host a fails to be taken over by the second network port
※ The first network port in the Host B is taken over by the second network when the fault occurs
※ Host a failure, Host B automatically take over the operation and data of host a
※ Host B takes over the host name of A and the network address (IP)
※ Host A's job will be run automatically on Host B
※ Host a client (client) can continue to run without having to log in again
※ Host A also can be used as B backup machine
※ Digital Duplex Backup and network card backup

Introduction to the new function of ROSE HA

Roseha for Windows

Increased support for Team type network cards.
Increased support for the hardware lock feature of the disk array.
Modified the configuration of the volume resource and configured it as a basic volume unit (disk).
Enhance the processing of shared volumes, when the host occurs when the volume resource switch is unsuccessful, the active shutdown or reboot, to ensure that the standby machine successfully.
For the NT Server type of resources, added a new configuration: In addition to the traditional direct selection of NT service way to configure, you can also specify the start, Stop, Agent three columns (program) to configure.
In the Userdefine type resource, the option to allow the agent to run the window is increased.
Enhanced editing of resources.
Increased the alarm mechanism used to share the disk character.
Increased ability to adapt to switch or hub power outages.
Enhanced processing of available shared resources for FileShare resources.
Adds the ability to save various configuration information to a text file in the Roseha interface.
Added content information for the log section.
The HostID mechanism has been redesigned to ensure that the system is reinstalled and the hard disk repartition does not change the HostID.
Other management interface in all aspects of operation and display modification.
The manual is adapted and supplemented for related functions.

Roseha for Linux

Supports a variety of mainstream and latest Linux operating systems, such as RedHat 7.x, RedHat 8.x.
New GUI management configuration interface based on JAVA2.
Support for the IANS capabilities of Intel network adapters, Intel network adapters must use Intel-supplied E100, e1000 drivers.
The configuration interface adds GroupID settings that make it easier to visually designate multiple services as a set of services that are switched in and out at the same time.
The configuration interface provides a way to configure multiple volumes, and you can configure more than one volume resource for a job.
Added the management interface modification function, can modify the job in the Bringout state directly in the management interface, need not remove and reconfigure.
The new disk system monitoring mechanism is added to enable more efficient, accurate and timely monitoring of disk array status.
Standardize the return value of the uniform agent, returning 0 values when normal, and not 0 values when abnormal.
Enhanced security to modify the partial execution of file permissions. Ensure that the root user can run the HA program.

Roseha for UnixWare

Support UnixWare 7.1.1 Latest patch Pack Uw711m2.ima
New GUI management configuration interface based on JAVA2.
The new disk system monitoring mechanism is added to enable more efficient, accurate and timely monitoring of disk array status.
Standardize the return value of the uniform agent, returning 0 values when normal, and not 0 values when abnormal.
Enhanced security to modify the partial execution of file permissions. Ensure that the root user can run the HA program.
Update the HostID mechanism. After installing Roseha, use the HostID tool provided by Roseha to obtain the HostID required for the application license and no longer use UnixWare as systemid as HostID. Ensure that the operating system is reinstalled and the hard disk repartition does not change HostID