CN103905253A - Server monitoring and management method based on Nagios and BMC - Google Patents

Abstract

The invention proposes a server monitoring and management method based on Nagios and BMC, which realizes Nagios' interactive access to server out-of-band monitoring information, expands the path of server monitoring and management, and improves the usability of the server. The invention mainly includes: server node installation and Nagios configuration, BMC-oriented Nagios extension plug-in implementation, IPMI protocol-based python script implementation, hardware communication interface implementation between Nagios and BMC, server remote monitoring management client and other parts. Through the information interaction between the Nagios tool and the BMC controller, the BMC-oriented function plug-in is expanded on the basis of the original Nagios-based in-band monitoring and management system, and the Nagios-oriented information acquisition is added on the basis of the original BMC-based out-of-band monitoring and management system , improve the availability of the server system through the key information interaction between the two systems.

Description

A server monitoring and management method based on Nagios and BMC

the

technical field

    The invention designs a server technology, specifically a server monitoring and management method based on Nagios and BMC.

Background technique

Servers play an important role in all aspects of today's society. Whether it is national defense, technology, finance and insurance, or banking, energy, and government enterprises, servers exist in almost all aspects. In order to ensure the stable operation of the above-mentioned tasks, it is a key premise to monitor and manage the server effectively in real time. There are many tools for server monitoring, such as open source tools Ganglia, Nagios, Zabbix, and professional tools Pingdom, interSeptorPro, Nimsoft, etc. These tools can monitor all aspects of server information, from running time, performance to security, and even where the server is located. physical environment index. Ganglia is used to measure thousands of server nodes, providing system static data and important performance measurement data, especially for cloud computing systems; Nagios is a server-level and network monitoring program that detects hosts and services, when abnormalities occur And the user can be reminded when it is dismissed. Zabbix is a source solution based on WEB interface that provides distributed system monitoring and network monitoring functions; Pingdom can monitor running time and overall performance, and generate easy-to-read tables and charts; interSeptor is an Ethernet data center and rack monitoring system, It can monitor the environmental conditions of computer rooms and racks, and issue early warnings in case of system failures or other situations that may endanger business continuity; NMS can monitor the core resources of servers and can centrally manage remote processes and services. Basically these tools allow you to see what problems might be going on with your server and fix them before they arise. A single open source tool has limited functions, requires a combination of multiple tools to play an effective role, and cannot get professional technical support, but can be obtained for free, while enterprise-level professional tools are powerful and equipped with professional technical support, but the price is very expensive .

In recent years, using open source tools to monitor and maintain server systems has become a mainstream solution. However, since these open source monitoring tools have their own characteristics and focuses, it is difficult to install all the tools in the actual monitoring and maintenance process. It is also difficult for the administrators who can install the system to have the energy to deal with the working conditions of all the tools, which brings great problems to the monitoring and management. Therefore, how to effectively integrate open source tool solutions, simplify the monitoring and management process, and comprehensively and accurately monitor and manage servers in server systems and even server cluster systems has become an urgent problem to be solved.

Contents of the invention

Aiming at how to effectively integrate open source tool solutions, simplify the monitoring and management process, and comprehensively and accurately monitor and manage the server system, the present invention proposes a server monitoring and management method based on Nagios and BMC.

The invention includes: server node installation and Nagios configuration, BMC-oriented Nagios extension plug-in realization, IPMI protocol-based python script realization, hardware communication interface realization between Nagios and BMC, server remote monitoring management client and other parts.

Among them, the installation and configuration of Nagios on the server node needs to be installed and configured on both the management and managed server nodes, so that Nagios can work normally, that is, the management node can obtain the monitoring and management information of the managed node through Nagios;

Among them, the Nagios extension plug-in facing BMC, the main purpose is to establish an interactive interface file between Nagios and the IPMI specification, and the Nagios official website can download the nagios_ipmi_monitor extension plug-in based on C language writing, the present invention utilizes the better cross-platform scripting language python to achieve, and On the original basis, the interface file for obtaining Nagios monitoring information for BMC is added.

1. The monitor_nagios_to_bmc_plugins plug-in group for Nagios to obtain BMC information mainly includes power supply, fan, voltage, temperature sensor status acquisition, energy management, log information export and other operations;

monitor_bmc_fan.py, monitor_bmc_temp.py,

monitor_bmc_psu.py, monitor_bmc_psu_control.py,

monitor_bmc_voltage.py, monitor_bmc_logs.py;

2. The get_nagios_to_bmc_plugins plug-in group for BMC to obtain Nagios information mainly includes status information such as CPU utilization, memory utilization, hard disk utilization, process status, and network performance indicators:

get_nagios_cpu_usgage.py, get_nagios_mem_usgage.py,

get_nagios_ps_status.py, get_nagios_hdd_status.py,

get_nagios_net_status.py;

Among them, the hardware communication interface between Nagios and BMC is the KCS interface based on the LPC protocol, which is suitable for the CPU of the Intel X86 platform;

Among them, the server remote monitoring and management client is installed in the Webserver of the BMC, and users can log in and access it through the WebBrowser;

Among them, the server remote monitoring management client is deployed on the BMC side, which has basic out-of-band monitoring and management functions, namely asset information detection, remote control maintenance, system log information, event alarms, etc. In addition, based on the key in-band information provided by Nagios, Added in-band information monitoring;

Among them, the present invention expands the BMC-oriented function plug-in on the basis of the original Nagios-based in-band monitoring and management system based on the information interaction between the Nagios tool and the BMC controller, and adds a oriented Nagios information acquisition;

The beneficial effects of the present invention are:

There is no need to install a specific Agent on the client server, but only need to install the BMC-oriented extension plug-in based on the standard IPMI protocol on the Nagios monitoring management tool that has already been installed by the client, which not only ensures the security of client information but also reduces the cost of system maintenance. difficulty.

 It enables customers to add information mutual backup to the original monitoring and management method, and can obtain key system information no matter from the in-band system or the out-of-band system, which increases the security and availability of system operation.

It can not only effectively integrate open source tool solutions, simplify the monitoring and management process, but also comprehensively and accurately monitor and manage the server system and the server in the system environment.

Description of drawings

Fig. 1 is a server monitoring and management architecture diagram based on Nagios and BMC according to the present invention;

Fig. 2 is a flow chart of implementing server monitoring and management according to the present invention.

Detailed ways

Embodiments of the present invention will be described below in conjunction with the accompanying drawings. It should be understood that the described embodiments are only used to illustrate and understand the present invention, and are not intended to limit the present invention.

Figure 1: It is a structure diagram of server monitoring and management based on Nagios and BMC according to the present invention. It mainly includes in-band monitoring and management tools, operating system OS layer, hardware platform layer, BMC layer, and out-of-band monitoring and management tools, as shown in Figure 1. The specific working process is described as follows:

1) The in-band monitoring management tool can be the Nagios server, which can be installed on the central monitoring server to obtain the alarm information sent by the Nagios client in real time;

2) Install Nagios on the local operating system of the managed node;

3) Configure the Nagios extension plug-in on the basis of installing the Nagios client, that is, use the monitor_nagios_to_bmc_plugins plug-in group for Nagios to obtain BMC information and the get_nagios_to_bmc_plugins plug-in group for BMC to obtain Nagios information. The specific implementation of the plug-in includes the following two parts, taking monitor_bmc_psu.py as an example ;

1. Nagios configuration file

Define service {

host name hostname

service_description power supply unit

 privilege                        

command monitor_bmc_pus!192.168.1.99!root!superuser

}

2. Specific python code

#!/usr/bin/python

from os import path,system

import sys, getpass

monitor_bmc_psu="/usr/lib/nagios/plugins/monitor_bmc_psu"

#logging's usage: logger.error(message),logger.info(message)

def initlog():

import logging

logger = logging. getLogger()

hdlr = logging. FileHandler(logfile)

formatter = logging.Formatter('%(asctime)s - %(levelname)s - %(message)s')

hdlr. setFormatter(formatter)

logger. addHandler(hdlr)

  logger.setLevel(logging.NOTSET)

return logger

logfile=__file__.replace(r'.py','.log')

logger = initlog()

#command execution

command_list=["psu_status","psu_poweron","psu_poweroff"]

def status(ip,password):

try:

system("echo ipmitool -I lan -H %s -U '%s' -P '%s' power status "%(ip,username,password))

except OSError, error:

logging. error(error)

def start(ip,password):

try:

system("echo ipmitool -I lan -H %s -U '%s' -P '%s' power on "%(ip,username,password))

except OSError, error:

logging. error(error)

def stop(ip,password):

try:

system("echo ipmitool -I lan -H %s -U '%s' -P '%s' power off "%(ip,username,password))

except OSError, error:

logging. error(error)

1) In this example, configure the KCS interface of the BMC and install the KCS driver in Linux;

2) The firmware program of BMC mainly supports the composition of IPMI stack, Web Server, module driver layer and interface layer

3) BMC obtains hardware asset information through SMBus bus;

4) The core module of the Nagios tool can obtain the hardware asset work information;

5) Out-of-band monitoring and management tools, mainly through WebBrowser to log in to BMC Web Server to obtain monitoring and management information;

Fig. 2: It is the implementation flowchart of server monitoring and management according to the present invention. As shown in Figure 2, the specific implementation steps are as follows:

1) Step 1: Install Nagios on the local operating system of the managed node;

2) Step 2: Configure Nagios extensions on the basis of installing the Nagios client;

3) Step 3: BMC configures KCS interface;

4) Step 4: Send commands to BMC through the KCS interface;

5) Step 5: The BMC side receives and parses the commands sent in-band;

6) Step 6: Determine what type of command it is?

7) Step 7: If it is the execution command To BMC type, then execute the corresponding IPMI Command to obtain hardware working status and other information;

8) Step 8: Send the information to the in-band system through the KCS interface;

9) Step 9: Nagios plugins parse and return information;

10) Step 10: Display out-of-band information on the local monitoring terminal;

11) Step 11: If it is an analysis command From Nagios type, analyze the information and judge whether it is necessary to generate an alarm message;

12) Step 12: If an alarm is required, send an alarm message to the remote monitoring management terminal in the form of SNMP Trap;

Using the server monitoring and management method based on Nagios and BMC in this embodiment, it can not only effectively integrate open source tool solutions, simplify the server system monitoring and management process, but also through the designed in-band and out-of-band monitoring and management information interaction methods, comprehensively Accurately monitor the real-time working status of the server system to enhance the availability and safety of the service system. In addition, the present invention also relates to a method for extending the Nagios plug-in in the Python script language, and the abstract design of the Python script can better realize the effectiveness of cross-platform implementation of the Nagios extension.

Claims (8)

1. A server monitoring and management method based on Nagios and BMC, characterized in that it comprises: Server monitoring management architecture and interface design based on Nagios and BMC; Implementation method of Nagios extension plug-in for BMC; The content of the python command script based on the IPMI protocol; Features of server remote monitoring and management client. 2. The server monitoring and management method based on Nagios and BMC according to claim 1, characterized in that, in point 1), in the management architecture and interface design, the architecture includes a server, an interactive interface layer, a baseboard controller BMC, and a remote Manage clients. 3. The server monitoring and management method based on Nagios and BMC according to claim 1, characterized in that, in point 1), in the management architecture and interface design, the interactive interface layer of the in-band and out-of-band management systems may include But not limited to I2C bus interface, KCS interface based on LPC protocol, CPU resource sharing, etc., the present invention uses the KCS interface of LPC. 4. The server monitoring and management method based on Nagios and BMC according to claim 1, characterized in that, in point 2), in the implementation method of the BMC-oriented Nagios extension plug-in, Nagios supports perl, shell, python and PHP, etc. The plug-in is written in language to expand the monitoring service. The plug-in extension for BMC is the function extension for the standard IPMI protocol, mainly including standard IPMI commands and third-party OEM IPMI commands. The present invention uses Python scripting language. 5. the server monitoring management method based on Nagios and BMC according to claim 4, is characterized in that, described BMC at least supports a kind of tool such as ipmitool, openIPMI and at least supports a kind of languages such as perl, shell, python and PHP. 6. The server monitoring and management method based on Nagios and BMC according to claim 1, characterized in that, in point 3), the python command script content of the IPMI protocol, that is, the server-side python command script, mainly completes the system in-band and out-of-band monitoring and management information interaction, including: (1) Nagios obtains BMC information: Nagios sends IPMI commands to BMC after parsing scripts to obtain server fan speed information, power supply status, sensor information such as temperature and voltage, system event log information, RAID controller information, etc., and sends the received The information is then parsed by the python command script and used for Nagios; (2) BMC obtains Nagios information: Nagios parses information such as system CPU and memory usage, disk utilization, and system process working status into IPMI commands through scripts and sends them to BMC, which then processes the information and sends it to the client. 7. The server monitoring and management method based on Nagios and BMC according to claim 1, characterized in that, in point 4), the features of the remote monitoring and management client of the server mainly include: (1) The client supports WEB access mode; (2) The host end of the client is the WebServer server of BMC; (3) The client supports standard IPMI protocol, SNMP protocol, etc.; (4) The client is divided into interface layer, analysis layer and storage layer, etc.; The interface layer is responsible for providing a communication interface with in-band Nagios and accepting commands sent by Nagios The parsing layer is responsible for parsing and executing the IPMI commands sent by Nagios The storage layer is responsible for saving the parsing or execution results of IPMI commands into the system cache (5) The client is suitable for a variety of different host types: Rack client: rack server BMC end Management client: two-tier management architecture, including the BMC side of the management node and the managed node Central client: a three-level or multi-level management architecture such as server clusters or cloud computing, that is, the BMC terminal of the central management node, sub-management nodes, and managed nodes. 8. The server monitoring and management method based on Nagios and BMC according to claim 1, characterized in that, in point 4), in the design method of remote monitoring and management client of the server, the purpose is to use Nagios and BMC in-band and out-of-band For information exchange, on the basis of the original Nagios-based in-band monitoring and management system, an out-of-band monitoring and management system based on BMC is added.

Images