CN105893261A - Method for automatically testing stability of BMC (baseboard management controller) of server - Google Patents

Abstract

The invention discloses a method for automatically testing the stability of a BMC (baseboard management controller) of a server, belonging to the field of server detection; the stability test under the BMC is realized by adopting the shell script, the rapid and continuous automatic detection of the BMC information of the plurality of servers can be realized, the test data is automatically recorded and sorted, and the visual test result is finally formed.

Description

A method for automatically checking the stability of server BMC

technical field

The invention discloses a method for automatically checking the stability of a server BMC, which belongs to the field of server testing.

Background technique

BMC: Baseboard Management Controller, motherboard management controller, supports the industry standard IPMI specification. This specification describes the management functions already built into the motherboard. These features include: local and remote diagnostics, console support, configuration management, hardware management and troubleshooting, and more.

IPMI, Intelligent Platform Management Interface, is a new generation of general interface standard that makes hardware management "intelligent". Users can use IPMI to monitor the physical characteristics of the server, such as temperature, voltage, fan status, power supply, and chassis intrusion. The biggest advantage of IPMI is that it is independent of BIOS and OS, so users can monitor the server BMC as long as the power is turned on, no matter whether it is on or off.

With the continuous development of IT technology, information services and increasingly powerful cloud computing services have higher and higher requirements for server stability. The server needs to stress test the product during the development stage and the production stage. BMC stability is an important test. Users will check the status of Fru, Lan, Log, and Sensor during the process of using the server, especially some Internet clients will continue to monitor the status of the BMC. Whether the status values of the server BMC are stable and reliable needs to be strictly checked during the test. However, there are more and more types of server research and development, and manual stress testing of multiple products at the same time consumes a lot of manpower and material resources, which is not conducive to improving efficiency. The invention provides a method for automatically checking the stability of the BMC of a server. The stability test under the BMC is realized by using a shell script, which can realize the rapid and continuous automatic detection of the BMC information of multiple servers, automatically record and organize the test data, and finally form a Intuitive test results, the process is completely automated, which simplifies the operation process, reduces test costs and manpower consumption, reduces unnecessary manual interaction processes, and greatly improves work efficiency.

Contents of the invention

Aiming at the problems in the prior art, the present invention provides a method for automatically checking the stability of the BMC of the server, and uses a shell script to realize the stability test under the BMC, which simplifies the operation process, reduces the test cost and manpower consumption, and greatly improves the work efficiency .

The concrete scheme that the present invention proposes is:

A method for automatically checking the stability of the server BMC:

Install the Linux kernel operating system and shell test script on the test machine, log in to the operating system as root, enter the BMC management IP of the inspection server, and obtain the BMC status after pinging;

Set related variables and script running parameters;

Create a result folder to record different BMC status results into different documents.

The shell test script sets a filter program to integrate and process the test results and only keep key data.

The testing machine inputs the BMC management IPs of multiple inspection servers, and obtains the respective BMC states after each ping is successful.

The different BMC state results include recording fru information results, recording chassis state information, recording lan information, recording mc information, and recording sensor information.

Create result folders with time and date names.

The benefits of the present invention are:

The test machine of the present invention installs the operating system and the shell test script of the linux kernel, logs in the operating system as root, inputs the BMC management IP of the inspection server, and obtains the BMC state after pinging; sets relevant variables and script operating parameters; sets up the result folder , record different BMC status results into different documents;

The present invention adopts the mode of shell script to automatically detect the stability of the server BMC, which can realize the synchronous detection of Fru, Lan, Log, CPU&PCH Sensor, etc., can realize the rapid and continuous automatic detection of the BMC information of multiple servers, and automatically record and organize them Test data, and finally form an intuitive test result.

Description of drawings

Fig. 1 schematic flow chart of the method of the present invention.

detailed description

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

Utilize the inventive method, concrete operation is:

Prepare the test machine and install the operating system based on the linux kernel. After the installation is complete, log in to the operating system as root; you can copy the test program and auxiliary test program to any folder under the system;

You can enter the BMC management IP and user name and password of a single or multiple servers, and obtain the BMC status after pinging;

echo -e "./bmcStress_remote.sh <ip> <user> <password>"

Run the program, set the relevant variables, and the number of script running cycles, and enter the number of cycles. You can use this parameter to verify the stability of each state of the BMC for a long time. A set of codes is as follows:

#STRESS_TIME: MINUTE (UNIT)

STRESS_TIME=2880

INTERVAL_TIME=5

LOOP=`echo ${STRESS_TIME}*60/${INTERVAL_TIME}|bc`

(3) Create a result folder with the name of time and date

RESULT_DIR=`date +%Y%m%d-%H%M`

if [ ! -d $RESULT_DIR ]

then

mkdir -p $RESULT_DIR

else

the fi

Record different BMC status results in different documents, combined with the above code, a set of code is as follows:

IPMI="/usr/bin/ipmitool"

function SENSOR_RECORD()

{

$IPMI fru >> $RESULT_DIR/fru.log &----read and record fru information

$IPMI chassis status >> $RESULT_DIR/chassis_status.log &----read and record chassis status information

$IPMI lan print 1 >> $RESULT_DIR/lan_print_1.log &----read and record lan information

$IPMI mc info >> $RESULT_DIR/mc_info.log &----read and record mc information

$IPMI sel list >> $RESULT_DIR/sel_list.log &----read and record log information

$IPMI sensor get CPU0_Temp >> $RESULT_DIR/CPU0_Temp.log &---read and record the information of each sensor

$IPMI sensor get CPU1_Temp >> $RESULT_DIR/CPU1_Temp.log &

$IPMI sensor get PCH_Temp >> $RESULT_DIR/PCH_Temp.log &

$IPMI sensor get Inlet_Temp >> $RESULT_DIR/Inlet_Temp.log &

$IPMI sensor get Outlet_Temp >> $RESULT_DIR/Outlet_Temp.log &

$IPMI sensor get SYS_12V >> $RESULT_DIR/SYS_12V.log &

$IPMI sensor get SYS_5V >> $RESULT_DIR/SYS_5V.log &

$IPMI sensor get SYS_3.3V >> $RESULT_DIR/SYS_3.3V.log &

$IPMI -b 0x6 -t 0x2c raw 0x2e 0xc8 0x57 0x01 0x00 0x01 0x1 0x0 |/usr/bin/xargs|awk '{print strtonum("0x"$4)}' >> $RESULT_DIR/pw_cpu.log &

$IPMI -b 0x6 -t 0x2c raw 0x2e 0xc8 0x57 0x01 0x00 0x01 0x2 0x0 |/usr/bin/xargs|awk '{print strtonum("0x"$4)}' >> $RESULT_DIR/pw_mem.log &

$IPMI sensor get Total_Power >> $RESULT_DIR/pw_total.log &

}

Set the filter program to complete the integration processing of the test results, and only keep the key data. Combined with the above code, a set of code is as follows:

function filter_sensor()

{

Awk '/Sensor Reading/{print $4}' $RESULT_DIR/CPU0_Temp.log >> $RESULT_DIR/CPU0_Temp.tmp

awk '/Sensor Reading/{print $4}' $RESULT_DIR/CPU1_Temp.log >> $RESULT_DIR/CPU1_Temp.tmp

awk '/Sensor Reading/{print $4}' $RESULT_DIR/PCH_Temp.log >> $RESULT_DIR/PCH_Temp.tmp

awk '/Sensor Reading/{print $4}' $RESULT_DIR/Inlet_Temp.log >> $RESULT_DIR/Inlet_Temp.tmp

awk '/Sensor Reading/{print $4}' $RESULT_DIR/Outlet_Temp.log >> $RESULT_DIR/Outlet_Temp.tmp

awk '/Sensor Reading/{print $4}' $RESULT_DIR/SYS_12V.log >> $RESULT_DIR/SYS_12V.tmp

awk '/Sensor Reading/{print $4}' $RESULT_DIR/SYS_5V.log >> $RESULT_DIR/SYS_5V.tmp

awk '/Sensor Reading/{print $4}' $RESULT_DIR/SYS_3.3V.log >> $RESULT_DIR/SYS_3.3V.tmp

awk '/Sensor Reading/{print $4}' $RESULT_DIR/pw_total.log >> $RESULT_DIR/pw_total.tmp

echo "CPU0_Temp, CPU1_Temp, PCH_Temp, Inlet_Temp, Outlet_Temp, SYS_12V, SYS_5V, SYS_3.3V, pw_cpu, pw_mem, pw_total" > $RESULT_DIR/Sensor_Total.csv

paste -d, $RESULT_DIR/CPU0_Temp.tmp $RESULT_DIR/CPU1_Temp.tmp $RESULT_DIR/PCH_Temp.tmp $RESULT_DIR/Inlet_Temp.tmp $RESULT_DIR/Outlet_Temp.tmp $RESULT_DIR/SYS_12V.tmp $RESULT_DIR/SYS_5V.tmp $RESULT_DIR .3V.tmp $RESULT_DIR/pw_cpu.log $RESULT_DIR/pw_mem.log $RESULT_DIR/pw_total.tmp >> $RESULT_DIR/Sensor_Total.csv

View the running results. After the test is completed, the results will be redirected to the folder generated under the test directory with the name of date and time. There will be a Sensor_Total.csv file under the folder, which records the test results of all tests.

The invention can monitor the status values of BMCs of multiple servers through scripts, and the returned results can be redirected to documents so as to check whether the data is normal. In the traditional way, we manually create various environments, and then check the BMC logs to check whether the status values are abnormal in different environments. Automated execution through scripts can reduce manual operations, save complete data, and improve test accuracy.

This test program can run under different products and different OS, and can monitor the BMC status of multiple servers at the same time. time, improving test efficiency.

Claims (5)

1. a method for Automated inspection server B MC stability, is characterized in that

Test machine installs operating system and the shell test script of linux kernel, in root identity logs to operating system, defeated BMC state is obtained after entering to check server B MC management IP, ping logical；

Correlated variables, and script operational factor are set；

Set up destination file folder, different BMC state outcome be recorded in different document.

Method the most according to claim 1, is characterized in that shell test script arranges filter program, to test result Integration process, only retain critical data.

Method the most according to claim 1 and 2, is characterized in that test machine input multiple stage inspection server B MC management IP, respectively Respective BMC state is obtained after ping is logical.

Method the most according to claim 3, is characterized in that different BMC state outcome includes recording fru information result, note Record chassis status information, record lan information, record mc information, record sensor information.

Method the most according to claim 4, it is characterized in that runs after fame with time and date sets up destination file folder.