CN105243012A - Linux based cluster network performance evaluating method - Google Patents

Linux based cluster network performance evaluating method Download PDF

Info

Publication number
CN105243012A
CN105243012A CN201510577369.3A CN201510577369A CN105243012A CN 105243012 A CN105243012 A CN 105243012A CN 201510577369 A CN201510577369 A CN 201510577369A CN 105243012 A CN105243012 A CN 105243012A
Authority
CN
China
Prior art keywords
cluster
network performance
network
linux
machine
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201510577369.3A
Other languages
Chinese (zh)
Inventor
刘云飞
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Inspur Electronic Information Industry Co Ltd
Original Assignee
Inspur Electronic Information Industry Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Inspur Electronic Information Industry Co Ltd filed Critical Inspur Electronic Information Industry Co Ltd
Priority to CN201510577369.3A priority Critical patent/CN105243012A/en
Publication of CN105243012A publication Critical patent/CN105243012A/en
Pending legal-status Critical Current

Links

Landscapes

  • Debugging And Monitoring (AREA)

Abstract

The present invention discloses a Linux based cluster network performance evaluating method. The method comprises the following implementation steps: establishing a cluster environment: according to actual application requirements, finishing establishment of the cluster environment, wherein the number of servers is at least two; installing an operating system and configuring system parameters: installing the operating system on all machines, and configuring the system parameters after finishing system installation; installing a network performance evaluating tool: installing the network performance evaluating tool on all the machines in a cluster, and configuring IP addresses of network ports that need to be evaluated; carrying out network performance evaluation: after finishing configuration, running an evaluation command to check network performance statuses of all nodes in the cluster; determining whether each node in the cluster has a problem or not according to a result, and performing targeted repairing. The implementation steps are implemented in a server installed with an operating system having a Linux kernel, and during operation, a user can log on by using a root identity. According to the method disclosed by the present invention, the network performance status of each node in the cluster can be checked, and whether the network performance of each node has a problem can be determined.

Description

A kind of cluster network performance estimating method based on linux
Technical field
The present invention relates to computer server technical field, be specifically related to a kind of cluster network performance estimating method based on linux.
Background technology
Along with the development of IT field technology, conventional informationization service and the increasingly powerful demand of cloud computing service to server become increasingly complex.As server product provider, the quality of a server, is whether can meet different application demands.Use due to single machine cannot meet the growth of the application demand of client, needs to build cluster to meet the demand of embody rule; In the case, the stability of cluster environment, performance are turned into the focus that client pays close attention to; In cluster environment, need to be communicated with by network, the stable and high-performance of network, stablizing of the whole cluster environment of guarantee.So be very necessary for network performance evaluation in cluster environment.
Server cluster just refers to a lot of server centered to get up and carries out same service together, similarly is only have a server in client.Cluster can utilize multiple computing machine to carry out parallel computation thus obtain very high computing velocity, also can backup with multiple computing machine, thus makes any one machine break whole system still normally to run.
GREENPLUM is a relevant database cluster, and it is actually the logical data base be combined into by several independently database service.Comprise the instrument that an instrument may be used for test database connectedness in this database, we can pass through this instrument, carry out the network performance evaluation between cluster; Also can be used for assessing the network stabilization of each node between cluster;
Shell script is similar to the batch processing under Windows/Dos, namely puts in advance in a file with all kinds of order, facilitates a program file of disposable execution, mainly facilitates keeper to carry out arranging or manage.But it is more powerful than the batch processing under Windows, than higher by the program efficiency of other programs editor,
Based on above-mentioned technology, if can shell script be passed through, simplify artificial operation, faulty operation also can be avoided to bring delay machine single-candidate problem.And by operation automatically, the scheme import operation of batch can be realized.
Summary of the invention
The technical problem to be solved in the present invention is: for above weak point, provide a kind of flexibly, be applicable to the network performance evaluation method of various server cluster environment.
The technical solution adopted in the present invention is:
Based on a cluster network performance estimating method of linux, described method performing step is as follows:
One, build cluster environment: according to the practical application request of oneself, cluster environment built, number of servers is minimum is 2;
Two, installing operating system parameter under configuration-system: by all machine installing operating systems, after system installation, configuration-system parameter;
Three, network performance evaluation instrument is installed: to all machines in cluster, network evaluation instrument is installed, and configure the IP address needing to assess network interface;
Four, carry out network performance evaluation: after having configured, run assessment order, check the network performance state of all nodes in cluster; Judge in cluster, whether each node has problems by result, and repair targetedly;
Described performing step completes in the server of operating system being provided with linux kernel, and during operation, user carries out with root identity logs.
Installing operating system and under configuration-system parametric procedure as follows:
According to different application scenarioss, different operating system is installed, system is installed completely, after system installation, parameter under adjustment System:
A, all machines to be renamed;
B, editor's machine hosts list; All machine lists are filled up in this file;
C, carry out arranging without code entry;
D, closedown iptables and selinux;
The systematic parameter of E, setting suggestion;
After F, in the cluster 1 machine has configured, copy configuration file to other machines and covered.
Network performance evaluation instrument process is installed as follows:
A, set up upload file folder;
B, upload greenplum-db-*.zip to a wherein machine;
C, decompress(ion);
D, installation;
E, all nodes are all pressed same operation, above-mentioned instrument is installed;
F, configure all need assessment the network interface network address;
G, set up GPhost list.
Described methods and results analytic process is as follows:
For 10,000,000,000 networks, if having lower than 1100MB/sec, existing problems are described, need adjustment System parameter, retest and see whether can revise;
If also cannot revise problem, analytic system CPU interrupts service condition, if interrupt very many, need adjust machine bios arrange and system under interruption optimize the content, reduce interruption, allow CPU have more ability process network traffics; After having adjusted, retest and check test result;
If result still has problems, possible node existing problems, change node or replacing relevant network components retests.
Described method long-time stability evaluation process is as follows:
When single test is no problem, carry out long-time repeatedly pressure test; Check its corresponding stability, check long operation, whether network performance is up to standard.
Beneficial effect of the present invention is:
The inventive method can between different server products critic network performance, can several operation systems be supported: redhat, centos, suse run application; The method is easy and simple to handle, has good ease for use; Clear in structure between each module in program, is convenient to safeguard; And can this network performance evaluation be passed through, find the in-problem place of network in cluster, and adjust accordingly, get rid of network to the impact of cluster application; The present invention uses SHELL script edit, can automatically perform for a long time, does not want execution manually, practical, applied widely, is easy to promote.
This method can be applied in various configuration, does not rely on hardware configuration, can flexibly, easily according to different application scene, carry out network performance evaluation, to meet different customer demands.In addition, this method is simple to operate, orders blunt general, for being unfamiliar with linux and needing the technician carrying out operating under linux platform to have reality and deep meaning.
Accompanying drawing explanation
Fig. 1 is the inventive method realization flow figure.
Embodiment
Below according to Figure of description, in conjunction with embodiment, the present invention is further described:
Embodiment 1:
As shown in Figure 1, a kind of cluster network performance estimating method based on linux, described method performing step is as follows:
One, build cluster environment: according to practical application request, cluster environment built, number of servers is minimum is 2; All machines are well placed, and connect the equipment such as all cables; All network interfaces used that needs are connected, and ensures that communication is normal; Powered on by all machines, check whether can there be display by normal boot-strap, all machinery and equipment are all normally detected;
Two, installing operating system parameter under configuration-system: by all machine installing operating systems, RHEL, SUSE series OS can be supported; After system installation, configuration-system parameter;
Three, network performance evaluation instrument is installed: to all machines in cluster, network evaluation instrument is installed; And configure the IP address needing to assess network interface;
Four, carry out network performance evaluation: after having configured, run assessment order, check the network performance state of all nodes in cluster; Judge in cluster, whether each node has problems by result, and repair targetedly;
Described performing step completes in the server of operating system being provided with linux kernel, and during operation, user carries out with root identity logs.
Cluster building completes, and relevant hardware device is just fixed up; First can carry out Performance Evaluation to the network portion in the running environment of entirety by this method, to guarantee that the network in whole cluster is no problem, the normal use of cluster can not be affected.
Embodiment 2:
On the basis of embodiment 1, the present embodiment installing operating system and under configuration-system parametric procedure as follows:
According to different application scenarioss, install different operating system, the inventive method supports the OS of RHEL, Centos, SUSE series; System is installed completely, after system installation, parameter under adjustment System;
A, all machines to be renamed: such as mdw, smdw, sdw1, sdw2
B, editor's machine hosts list: vi/etc/hosts; All machine lists are filled up in this file; Such as: 192.168.1.11mdw
C, carry out arranging without code entry: ssh-keygen-trsa;
D, closedown iptables and selinux
cluster_run_all_nodes"hostname;serviceiptablesstop"
cluster_run_all_nodes"hostname;chkconfigiptablesoff"
cluster_run_all_nodes"hostname;chkconfigip6tablesoff"
cluster_run_all_nodes"hostname;chkconfiglibvirtdoff"
cluster_run_all_nodes"hostname;setenforce0"
cluster_run_all_nodes"hostname;sestatus"
vi/etc/selinux/config
cluster_copy_all_nodes/etc/selinux/config/etc/selinux/
The systematic parameter of E, setting suggestion
vi/etc/sysctl.conf
kernel.shmmax=500000000
kernel.shmmni=4096
kernel.shmall=4000000000
kernel.sem=2505120001002048
kernel.sysrq=1
kernel.core_uses_pid=1
kernel.msgmnb=65536
kernel.msgmax=65536
net.ipv4.tcp_syncookies=1
net.ipv4.ip_forward=0
net.ipv4.conf.default.accept_source_route=0
net.ipv4.tcp_tw_recycle=1
net.ipv4.tcp_max_syn_backlog=4096
net.ipv4.conf.all.arp_filter=1
net.ipv4.conf.default.arp_filter=1
net.core.netdev_max_backlog=10000
vm.overcommit_memory=2
kernel.msgmni=2048
net.ipv4.ip_local_port_range=102565535
vi/etc/security/limits.conf
*softnofile65536
*hardnofile65536
*softnproc131072
*hardnproc131072
After F, in the cluster 1 machine has configured, copy configuration file to other machines and covered.
Embodiment 3:
On the basis of embodiment 1, the present embodiment installation network performance evaluation instrument process is as follows:
A, set up upload file folder
mkdir-p/tmp/soft
B, upload greenplum-db-*.zip to a wherein machine
cpgreenplum-db-*.zip/tmp/soft
cd/tmp/soft
C, decompress(ion)
unzipgreenplum-db-*.zip
D, installation
/bin/bashgreenplum-db-*.bin
E, all nodes are all pressed same operation, above-mentioned instrument is installed;
F, configure all need assessment the network interface network address, on a machine during multiple network interface, be not arranged on the same network segment; Such as below shown in content:
Configuration/etc/hosts
192.168.9.123mdw
172.16.10.101mdw-1
172.16.11.101mdw-2
192.168.9.124smdw
172.16.10.102smdw-1
172.16.11.102smdw-2
192.168.9.125sdw1
172.16.10.1sdw1-1
172.16.11.1sdw1-2
G, set up GPhost list.
Embodiment 3:
On the basis of embodiment 1, the present embodiment execution performance assessment command procedure is as follows:
gpcheckperf-fnetperf1-4-rN-d/tmp
As follows for ten thousand Broadcom Output rusults:
-------------------
--NETPERFTEST
-------------------
====================
==RESULT
====================
Netperfbisectionbandwidthtest
mdw-3->smdw-3=1118.480000
sdw1-3->mdw-3=1118.500000
smdw-3->mdw-3=1118.270000
mdw-3->sdw1-3=1118.580000
Summary:
sum=4473.83MB/sec
min=1118.27MB/sec
max=1118.58MB/sec
avg=1118.46MB/sec
median=1118.50MB/sec
The above results shows that network is no problem.
Embodiment 4:
On the basis of above-described embodiment, described in the present embodiment, methods and results analytic process is as follows:
For 10,000,000,000 networks, if having lower than 1100MB/sec, existing problems are described, need adjustment System parameter, retest and see whether can revise;
If also cannot revise problem, analytic system CPU interrupts service condition, if interrupt very many, need adjust machine bios arrange and system under interruption optimize the content, reduce interruption, allow CPU have more ability process network traffics; After having adjusted, retest and check test result;
If result still has problems, possible node existing problems, change node or replacing relevant network components retests.
Embodiment 5:
On the basis of embodiment 4, described in the present embodiment, method long-time stability evaluation process is as follows:
When single test is no problem, long-time repeatedly pressure test can be carried out; Check its corresponding stability, check long operation, whether network performance is up to standard.
Above embodiment is only for illustration of the present invention; and be not limitation of the present invention; the those of ordinary skill of relevant technical field; without departing from the spirit and scope of the present invention; can also make a variety of changes and modification; therefore all equivalent technical schemes also belong to category of the present invention, and scope of patent protection of the present invention should be defined by the claims.

Claims (5)

1. based on a cluster network performance estimating method of linux, it is characterized in that, described method performing step is as follows:
One, build cluster environment: according to the practical application request of oneself, cluster environment built, number of servers is minimum is 2;
Two, installing operating system parameter under configuration-system: by all machine installing operating systems, after system installation, configuration-system parameter;
Three, network performance evaluation instrument is installed: to all machines in cluster, network evaluation instrument is installed, and configure the IP address needing to assess network interface;
Four, carry out network performance evaluation: after having configured, run assessment order, check the network performance state of all nodes in cluster; Judge in cluster, whether each node has problems by result, and repair targetedly;
Described performing step completes in the server of operating system being provided with linux kernel, and during operation, user carries out with root identity logs.
2. a kind of cluster network performance estimating method based on linux according to claim 1, is characterized in that, installing operating system and under configuration-system parametric procedure as follows:
According to different application scenarioss, different operating system is installed, system is installed completely, after system installation, parameter under adjustment System:
A, all machines to be renamed;
B, editor's machine hosts list; All machine lists are filled up in this file;
C, carry out arranging without code entry;
D, closedown iptables and selinux;
The systematic parameter of E, setting suggestion;
After F, in the cluster 1 machine has configured, copy configuration file to other machines and covered.
3. a kind of cluster network performance estimating method based on linux according to claim 1, is characterized in that, installation network performance evaluation instrument process is as follows:
A, set up upload file folder;
B, upload greenplum-db-*.zip to a wherein machine;
C, decompress(ion);
D, installation;
E, all nodes are all pressed same operation, above-mentioned instrument is installed;
F, configure all need assessment the network interface network address;
G, set up GPhost list.
4., according to the arbitrary described a kind of cluster network performance estimating method based on linux of claim 1,2 or 3, it is characterized in that, described methods and results analytic process is as follows:
For 10,000,000,000 networks, if having lower than 1100MB/sec, existing problems are described, need adjustment System parameter, retest and see whether can revise;
If also cannot revise problem, analytic system CPU interrupts service condition, if interrupt very many, need adjust machine bios arrange and system under interruption optimize the content, reduce interruption, allow CPU have more ability process network traffics; After having adjusted, retest and check test result;
If result still has problems, possible node existing problems, change node or replacing relevant network components retests.
5. a kind of cluster network performance estimating method based on linux according to claim 4, is characterized in that, described method long-time stability evaluation process is as follows:
When single test is no problem, carry out long-time repeatedly pressure test; Check its corresponding stability, check long operation, whether network performance is up to standard.
CN201510577369.3A 2015-09-11 2015-09-11 Linux based cluster network performance evaluating method Pending CN105243012A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510577369.3A CN105243012A (en) 2015-09-11 2015-09-11 Linux based cluster network performance evaluating method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510577369.3A CN105243012A (en) 2015-09-11 2015-09-11 Linux based cluster network performance evaluating method

Publications (1)

Publication Number Publication Date
CN105243012A true CN105243012A (en) 2016-01-13

Family

ID=55040665

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510577369.3A Pending CN105243012A (en) 2015-09-11 2015-09-11 Linux based cluster network performance evaluating method

Country Status (1)

Country Link
CN (1) CN105243012A (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105808433A (en) * 2016-03-04 2016-07-27 浪潮软件集团有限公司 One-key tuning method based on Linux operating system
CN106161144A (en) * 2016-07-26 2016-11-23 浪潮电子信息产业股份有限公司 The method of test Smart Rack equipment cabinet server network performance
CN108196992A (en) * 2017-12-28 2018-06-22 广州品唯软件有限公司 A kind of automatic performance test method and device
CN108833168A (en) * 2018-06-20 2018-11-16 郑州云海信息技术有限公司 A kind of server cluster environment network system and Detection of Stability method
CN110798338A (en) * 2019-09-30 2020-02-14 烽火通信科技股份有限公司 Edge cloud deployment method and device
CN112286538A (en) * 2020-10-28 2021-01-29 电信科学技术第十研究所有限公司 Greenplus database automatic installation method and device

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030009511A1 (en) * 2001-07-05 2003-01-09 Paul Giotta Method for ensuring operation during node failures and network partitions in a clustered message passing server
CN104063265A (en) * 2014-07-04 2014-09-24 云南电网公司 Method for comprehensively evaluating virtual resources
CN104579853A (en) * 2015-02-10 2015-04-29 浪潮电子信息产业股份有限公司 Method for network testing of server cluster system

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030009511A1 (en) * 2001-07-05 2003-01-09 Paul Giotta Method for ensuring operation during node failures and network partitions in a clustered message passing server
CN104063265A (en) * 2014-07-04 2014-09-24 云南电网公司 Method for comprehensively evaluating virtual resources
CN104579853A (en) * 2015-02-10 2015-04-29 浪潮电子信息产业股份有限公司 Method for network testing of server cluster system

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105808433A (en) * 2016-03-04 2016-07-27 浪潮软件集团有限公司 One-key tuning method based on Linux operating system
CN106161144A (en) * 2016-07-26 2016-11-23 浪潮电子信息产业股份有限公司 The method of test Smart Rack equipment cabinet server network performance
CN108196992A (en) * 2017-12-28 2018-06-22 广州品唯软件有限公司 A kind of automatic performance test method and device
CN108196992B (en) * 2017-12-28 2021-05-28 广州品唯软件有限公司 Automatic performance testing method and device
CN108833168A (en) * 2018-06-20 2018-11-16 郑州云海信息技术有限公司 A kind of server cluster environment network system and Detection of Stability method
CN108833168B (en) * 2018-06-20 2021-08-20 郑州云海信息技术有限公司 Server cluster environment network system and stability detection method
CN110798338A (en) * 2019-09-30 2020-02-14 烽火通信科技股份有限公司 Edge cloud deployment method and device
CN112286538A (en) * 2020-10-28 2021-01-29 电信科学技术第十研究所有限公司 Greenplus database automatic installation method and device
CN112286538B (en) * 2020-10-28 2023-02-17 电信科学技术第十研究所有限公司 Greenplus database automatic installation method and device

Similar Documents

Publication Publication Date Title
US11038784B2 (en) Techniques for evaluating server system reliability, vulnerability and component compatibility using crowdsourced server and vulnerability data
CN105243012A (en) Linux based cluster network performance evaluating method
US9923867B2 (en) Security threat identification, isolation, and repairing in a network
US9811443B2 (en) Dynamic trace level control
US20150277943A1 (en) Policy enforcement in a virtualized environment
US8990372B2 (en) Operation managing device and operation management method
US9535727B1 (en) Identifying virtual machines that perform inconsistent with a profile
CN106919485B (en) System based on hardware testing tool configured on server
US9086942B2 (en) Software discovery by an installer controller
US10178003B2 (en) Instance based management and control for VM platforms in virtual processing environments
US10614218B2 (en) Scan time reduction in application code security scanning
CN103970655A (en) Server automatic cluster testing method based on expect
US11799875B2 (en) Computerized system for complying with certain critical infrastructure protection requirements
KR101594701B1 (en) Apparatus and method for detecting abnormal connection
US20160219104A1 (en) Automated Assessment Report Generation
CN110798341A (en) Service opening method, device and system
US20170017602A1 (en) Storage system cabling analysis
US11599404B2 (en) Correlation-based multi-source problem diagnosis
CN104618196A (en) Method for identifying compatibility of server and Citrix system
US11411819B2 (en) Automatic network configuration in data protection operations
US11487570B1 (en) Efficient creation of endpoints for accessing services directly within a cloud-based system
CN108021407B (en) Service processing method and device based on network equipment
US20140089482A1 (en) Device management for determining the affects of management actions
CN113556252B (en) Method and system for checking and repairing network equipment baseline configuration
US11178110B2 (en) Controlling compliance remediations

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20160113