CN101404568A - Double-network card hot backup redundancy method - Google Patents

Abstract

The invention relates to a dual network card hot backup redundant method, comprising: connecting two network cards of each machine to two switches in the network topology; cascading the two switches; using the network card monitoring program to utilize the working network card The IP address of the switch sends a message to the switch and receives a response. When multiple consecutive messages do not receive a response after a certain time interval, it is judged that the working network card or network device is faulty. The system sends out an alarm signal and records the alarm status. Switch the IP address to the backup network card; when the working network card returns to normal, switch the original IP address back to the working network card. The present invention can realize automatic switching when a working network card or network equipment or switch fails; and when the working network card returns to normal, it can also be switched back, the switching program is automatically carried out, and fault alarms and records can be carried out, which is a good solution cross-fault problem.

Description

Dual NIC Hot Standby Redundancy Method

technical field

The invention relates to a dual-network card hot standby redundancy method for a server, which is used for monitoring the state of network communication equipment and communication management, and belongs to the technical field of communication management.

Background technique

Today's servers generally have two or more network cards, and some operating systems such as IBM and SUN can use their own software to aggregate the network cards to achieve the purpose of hot standby redundancy. These aggregation software are part of the operating system and depend on the support of the operating system, and some software also depend on the support of the switch.

Contents of the invention

In order to solve the deficiencies in the prior art, the purpose of the present invention is to provide a dual-network card hot standby redundancy method with high reliability and high availability at the application layer, which does not depend on specific operating systems and switches, and is applicable to various UNIX and Linux operating systems. The present invention is responsible for the state monitoring and communication management of the network communication equipment, and guarantees the normal operation of the system and no data loss when any network card equipment is still working normally, so as to achieve the purpose of ensuring the reliability of network communication.

In order to achieve the above-mentioned purpose of the invention, the present invention is achieved by adopting the following technical solutions:

The function realized by the invention includes two parts: (1) network card fault detection function and (2) hot standby redundant function of double network cards.

Technical solution of the present invention: a dual network card hot standby redundant method is characterized in that comprising the following steps:

(1) In the network topology, connect the two network cards of each machine to two switches;

(2) Cascading two switches to each other in the network topology;

(3), the IP address of the network card monitoring program is used to send messages to the switch and receive responses. When multiple consecutive messages are not answered after a certain time interval, it is judged that the network card or network equipment is faulty. The system sends out an alarm signal and records the alarm status, and switches the IP address to the backup network card;

(4) When the working network card returns to normal, switch the original IP address back to the working network card.

Aforesaid dual-card hot standby redundant method is characterized in that in above-mentioned step 3, configures fixed IP and service IP on each network card, and described fixed IP is used for network card monitoring program to send message to monitor network card state Address, the service IP is the address provided to the system when the system works normally, and the service IP on the two network cards can be provided to the system at the same time; carry out the following detection switching process:

(1) judge whether the service IP is on the working network card by the network card monitoring program;

(2) If the service IP is working on the working network card, it will not be switched, and cyclic monitoring will be carried out; when the working network card or network equipment is not working properly, and multiple consecutive messages have not been answered after a certain time interval, check The working status of the backup network card; if the backup network card is working normally, switch the service IP to run on the backup network card.

Aforesaid dual network card hot standby redundant method is characterized in that in above-mentioned step 4, service IP works on the backup network card, and the network card monitoring program constantly judges the state of the work network card, if the work network card is normal, judges the backup network card again; If the backup network card is also normal, wait for the working network card to maintain a normal state for a period of time before switching the service IP to the working network card; if the backup network card is not normal, immediately switch the service IP to the working network card.

The aforementioned dual-card hot standby redundancy method is characterized in that the service IP and the fixed IP are not on the same network segment.

Advantages of the present invention: 1. Network failure monitoring and failure network card switching are all completed by automatic programs, without manual intervention; 2. It is suitable for various UNIX operating systems and LINUX operating systems, and has good versatility; 3. All functions are completely in the The application layer is implemented without the help of any software that comes with the operating system; 4. The switching speed after the network failure is fast, which can reach within seconds; 5. The parameters such as the fault detection cycle and the connection switching time interval are programmable and adjustable; 6. Fixed IP Not in the same network address segment as the service IP, the fixed IP will not participate in the general business of the system, and will not conflict with the service IP of other machines; 7. There are service IPs on both network cards, which can be provided to the system at the same time; 8. It can ensure that the entire system can still communicate normally regardless of the failure of the network card of the host, the failure of the network cable, or the failure of a switch, which solves the cross-fault problem of traditional dual-network technology.

Description of drawings

Fig. 1 is a flow chart of network card fault detection processing;

Fig. 2 is the network topology diagram of dual-network hot standby redundancy method;

Figure 3 is a schematic diagram of "cross fault" and "single point of failure" in traditional dual network technology;

FIG. 4 is a flow chart of service IP switching processing.

Detailed ways

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

The dual-network card hot standby redundancy method involved in the present invention is realized by self-developed software working on the application layer, does not depend on specific operating systems and switches, and is applicable to various UNIX and Linux operating systems. The dual-card hot backup redundant method of the present invention has guaranteed portability, and can run safely and stably on operating system platforms such as IBM AIX, Sun Solaris, Alpha Tru64, HPUX, Linux; the system programming language adopts ANSI C/C++, And complies with the IEEE POSIX.2 standard.

In the present invention, the two network cards of each machine are respectively connected to two switches in the network topology, and the two switches are cascaded to each other in the network topology. The present invention first realizes the fault detection function of the network card, and detects the connectivity between the host and the connected network equipment (such as a switch). The functions it has are: (1) monitoring whether the network card and its auxiliary equipment (such as network cable, etc.) are working normally; (2) if the network card and its auxiliary equipment (such as network cable, etc.) When its auxiliary equipment (such as network cable, etc.) fails, record the fault status, and send an alarm signal to notify the responsible personnel to repair the network card fault.

Network card failure detection is realized by using the network card monitoring program to send ICMP messages and receive responses. A configuration file is set up in the system, and the address of the network device connected to each network card is configured. The automation program implemented by the system programming language runs in the form of a daemon process, reads the information of the configuration file, and monitors the connection status of the network card by continuously sending ICMP messages to the address of the network device connected to each network card. When multiple consecutive ICMP messages are not answered after a certain time interval, it will be judged that the network device is faulty. The processing flow is shown in Figure 1.

The second part of the functions realized by the present invention is the hot standby redundancy mechanism of dual network cards. The network topology of the dual-network hot standby redundancy mechanism is shown in Figure 2. Compared with the traditional dual-network mechanism, the two switches in Figure 2 are cascaded. Through the cascading of switches, the problems of "cross fault" (Figure 3(a)) and "single point of failure" (Figure 3(b)) brought about by traditional dual-network technology can be solved. In Figure 3(a), when host 1 network card 1/network cable and host 2 network card 2/network cable fail, host 1 and host 2 cannot communicate; similarly, in Figure 3(b), when host 1 network card 1 / When the network cable fails, host 1 and host 2 cannot communicate. And the present invention can solve this problem well.

The hot standby redundancy mechanism of the dual network card is realized by configuring fixed IP and floating IP on the network card, and writing the above two addresses in the configuration file at the same time, and running the network card monitoring program and the network card switching program at the same time. Floating IP, also known as service IP, is the IP used for communication when the system works normally. The fixed IP and service IP are first configured on a designated network card (called the working network card of the service IP), the network card monitoring program uses the fixed IP to send ICMP messages to monitor the status of the network card, and the service IP is provided for normal use of the system. And the service IP and the fixed IP are not on the same network segment, so the fixed IP will not occupy the number of service IPs, and it can be guaranteed that the fixed IP will not be used by the general business of the system.

When the working network card of the service IP and its auxiliary equipment (such as network cables, etc.) fail, the network card switching program is responsible for switching the service IP to another network card (called the backup network card of the service IP) that is working in a normal state, but fixed IP does not move. After the working network card of the service IP returns to normal for a period of time (such as 3 seconds), the network card switching program is responsible for switching the service IP back to its working network card. The processing flow is shown in Figure 4.

Referring to accompanying drawing 4, at first judge whether service IP is on the working network card by the network card monitoring program.

If the service IP is working on the working network card, when the network card is not working properly, check the working status of the backup network card. If the backup network card works normally, the network card switching program will switch the service IP to the backup network card to run.

If the service IP is already working on the backup network card, it means that a switch from the working network card to the backup network card has occurred. Then the network card monitoring program judges the working network card status of the service IP. If the working network card is normal, then judge the status of the backup network card. If the backup network card is also normal, wait for the working network card to maintain a normal state for a period of time, and then switch the service IP to the working network card after the network card switching program. If the backup network card is also abnormal, immediately switch the service IP to the working network card.

working example

Referring to the accompanying drawing 2, the network card 1 of the host 1 is set with a floating IP: 192.168.0.1, and a fixed IP: 192.168.6.1. Network card 2 of host 1 sets floating IP: 192.168.1.1, and sets fixed IP: 192.168.7.1.

Network card 1 of host 2 sets floating IP: 192.168.0.2, and sets fixed IP: 192.168.6.2. Network card 2 of host 2 sets floating IP: 192.168.1.2, and sets fixed IP: 192.168.7.2.

The address of switch 1 is set to 192.168.6.254.

The address of switch 2 is set to 192.168.7.254.

The network card monitoring program uses the fixed IP of network card 1: 192.168.6.1 to send ICMP messages to the address 192.168.6.254 of switch 1. If you can get a response within a certain period of time, it is considered that the network status is normal. If no response is received within a certain time interval, it is considered that the network status is abnormal, and the network card switching program switches the floating IP of network card 1: 192.168.0.1 to network card 2. At this time, network card 2 has three ip addresses:

192.168.0.2

192.168.1.2

192.168.0.1

There are two types of network failures at this time. The first case is that the network card 1 of host 1 is faulty or the network cable is faulty. In this case, only the service IP of network card 1 of host 1 is switched to network card 2. Other machines can still communicate with the service IP of NIC 1 on NIC 2 through switch 1. The network communication of the whole system is normal. The second case is that switch 1 fails. In this case, the service addresses of network card 1 of all hosts are switched to network card 2. The entire system uses switch 2 for communication, and the network communication of the entire system is normal.

When the floating IP of network card 1 of host 1: 192.168.0.1 is switched to network card 2, the network card monitoring program still uses the fixed IP of network card 1: 192.168.6.1 to send ICMP messages to the address of switch 1: 192.168.6.254. If the response is received again, and the response is received for a continuous period of time, it is considered that the network card 1 is back to normal. At this time, the switching program of network card 1 switches the floating address of network card 1 from network card 2 back to network card 1. The entire system network communication is still normal.

The above embodiments do not limit the present invention in any form, and all technical solutions obtained by means of equivalent replacement or equivalent transformation fall within the protection scope of the present invention.

Claims (4)

1, a kind of double-network card hot backup redundancy method is characterized in that may further comprise the steps:

(1), in network topology, two network interface cards of every machine is connected on respectively on two switches;

(2), in network topology with two mutual cascades of switch;

(3), utilizing the IP address of work network interface card to send message and accept to switch by the network interface card supervisory programme replys, when a plurality of continuous messages are not replied after certain time interval, then be judged as work network interface card or network equipment failure, system sends alarm signal and writes down alarm status, and the IP address switchover is arrived the backup network interface card;

(4), recover just often former IP address switchover to be returned the work network interface card when the work network interface card.

2, two card hot backup redundancy method according to claim 1, it is characterized in that in above-mentioned step 3, on each network interface card, dispose fixedly IP and service IP, described fixedly IP is used for the network interface card supervisory programme to send the address that message monitors network card status, described service IP offers the address that system uses when being system's operate as normal, the service IP on two network interface cards can offer system simultaneously and use; Carry out following detection handoff procedure:

(1) judges that by the network interface card supervisory programme service IP is whether on the work network interface card;

(2) if service IP is operated on the work network interface card, then do not switch, supervision circulates; Undesired when this work network interface card or network equipment work, when a plurality of continuous messages do not obtain replying, check the operating state of backup network interface card after certain time interval; If the backup network interface card is working properly, then service IP is switched on the backup network interface card and move.

3, double-network card hot backup redundancy method according to claim 2 is characterized in that in above-mentioned step 4, and service IP is operated on the backup network interface card, and the state of the continuous judgment task network interface card of network interface card supervisory programme if the work network interface card is normal, is judged the backup network interface card again; If the backup network interface card is also normal, wait for that then the work network interface card keeps normal condition and after a period of time service IP switched to the work network interface card; If the backup network interface card is undesired, then immediately service IP is switched to the work network interface card.

4, according to claim 2 or 3 described two card hot backup redundancy methods, it is characterized in that described service IP and fixedly IP not on the same network segment.

Images