Basic Optimization after CentOS6.4 system installation

Source: Internet
Author: User
Tags nameserver rsyslog

During the O & M work, we found that the Linux system cannot be immediately put into the production environment after installation, which usually requires optimization by our O & M personnel.

Next, I will briefly explain some basic optimization operations on Linux after installation.

Note: This optimization is based on CentOS (5.8/6.4 ). I will mention the minor differences between 5.8 and 6.4 During optimization.


Optimization entry:

Modify the IP address, gateway, host name, DNS, and so on to disable selinux, clear iptables, add common users, perform sudo authorization management, update yum source, and install necessary software, and automatically update server time. Streamline startup, and automatically clear/var/spool/clientmqueue/directory junk files at startup, when inode nodes are fully occupied, the default ssh service port is changed. the root user is prohibited from remotely locking the key file system and adjusting the file descriptor size and character set, enable it to remove Chinese characters and display Kernel Parameter Optimization on the screen before kernel version Logon


1. Modify the IP address, gateway, host name, DNS, etc.

[Root @ localhost ~] # Vi/etc/sysconfig/network-scripts/ifcfg-eth0DEVICE = eth0 # Nic name BOOTPROTO = static # static IP address acquisition status such: DHCP indicates Automatically Obtaining the IP address IPADDR = 192.168.1.113 # IP address NETMASK = 255.255.255.0 # subnet mask ONBOOT = yes # enable GATEWAY = 192.168.1.1 during boot [root @ localhost ~] # Cat/etc/sysconfig/network-scripts/ifcfg-eth0DEVICE = eth0BOOTPROTO = staticIPADDR = 192.168.1.113NETMASK = 255.255.255.0ONBOOT = yesGATEWAY = 192.168.1.1 [root @ localhost ~] # Vi/etc/sysconfig/networkHOSTNAME = c64 # modify the Host Name and restart GATEWAY = 192.168.1.1 # modify the default GATEWAY. If no GATEWAY is configured in eth0, the gateway here is used by default. [Root @ localhost ~] # Cat/etc/sysconfig/networkHOSTNAME = c64GATEWAY = 192.168.1.1 we can also use hostnamec64 to temporarily modify the Host Name and log on again to modify DNS [root @ localhost ~] # Vi/etc/resolv. conf # modify DNS information nameserver 114.114.114.114nameserver 8.8.8.8 [root @ localhost ~] # Cat/etc/resolv. conf # view the modified DNS information nameserver 114.114.114.114nameserver 8.8.8.8 [root @ localhost ~] # Service network restart # restart the NIC to take effect. You can also run the following command [root @ localhost ~] #/Etc/init. d/network restart


2. Disable selinux and clear iptables.

Disable selinux

[Root @ c64 ~] # Sed-I's/SELINUX = enforcing/SELINUX = disabled/G'/etc/selinux/config # modifying the configuration file takes effect permanently, but the system must be restarted. [Root @ c64 ~] # Grep SELINUX = disabled/etc/selinux/configSELINUX = disabled # view the changed result [root @ c64 ~] # Setenforce 0 # temporary effective command [root @ c64 ~] # Getenforce # view the current state of selinux: Permissive


Clear iptables

[Root @ c64 ~] # Iptables-F # clear firewall rules [root @ c64 ~] # Iptables-L # view firewall rule Chain INPUT (policy ACCEPT) target prot opt source destination Chain FORWARD (policy ACCEPT) target prot opt source destination Chain OUTPUT (policy ACCEPT) target prot opt source destination [root @ c64 ~] #/Etc/init. d/iptables save # save firewall configuration information


3. Add and executesudoAuthorization management

[Root @ c64 ~] # Useradd sunsky [root @ c64 ~] # Echo "123456" | passwd -- stdin sunsky & history-c [root @ c64 ~] # Mongodo Add the following content under root ALL = (ALL) ALL in this row: sunsky ALL = (ALL) ALL


4. Update yum source and necessary software installation

Yum install software. By default, the rpm package is obtained from the foreign official source and changed to the domestic source.

Two fast domestic sites: Sohu image site and Netease image site

Method 1: configure the source configuration file and upload it to linux.

Method 2: Install the source configuration file using the yum configured on the Image site

[root@c64 ~]# cd /etc/yum.repos.d/[root@c64 yum.repos.d]# /bin/mv CentOS-Base.repo CentOS-Base.repo.bak[root@c64 yum.repos.d]# wget http://mirrors.163.com/.help/CentOS6-Base-163.repo


Run the following command to check whether yum is normal.

[Root @ c64 yum. repos. d] # yum clean all # Clear yum cache [root @ c64 yum. repos. d] # yum makecache # create yum Cache


Run the following command to update the system to the latest version.

[Root @ c64 yum. repos. d] # rpm -- import/etc/pki/rpm-gpg/RPM-GPG-KEY * # import the signature KEY to RPM [root @ c64 yum. repos. d] # yum upgrade-y # update the system kernel to the latest version

Next we need to install several necessary software.

[root@c64 yum.repos.d]# yum install lrzsz ntpdate sysstat -y


Lrzsz is a software for uploading and downloading data.

Ntpdate is a software used to update the time of a remote time server.

Sysstat is a tool used to detect system performance and efficiency.


5. automatically update server time on a regular basis

[root@c64 ~]# echo '*/5 * * * * /usr/sbin/ntpdate time.windows.com >/dev/null 2 >&1' >>/var/spool/cron/root[root@c64 ~]# echo '*/10 * * * * /usr/sbin/ntpdate time.nist.gov >/dev/null 2>&1' >>/var/spool/cron/root


Tip: The time synchronization command paths for CentOS 6.4 are different.

6 is/usr/sbin/ntpdate

5 is/sbin/ntpdate

Extension: when the number of machines is small, the synchronization time of the above scheduled tasks is enough. If the number of machines is large, you can deploy another time synchronization Server NTP Server in the network. This is only mentioned here and is not deployed.

Time synchronization server architecture diagram:


6. Streamline boot self-starting services

After the operating system is installed, only crond, network, syslog, and sshd services can be retained. (Centos6.4 is rsyslog)

[root@c64 ~]# for sun in `chkconfig --list|grep 3:on|awk '{print $1}'`;do chkconfig --level 3 $sun off;done[root@c64 ~]# for sun in crond rsyslog sshd network;do chkconfig --level 3 $sun on;done[root@c64 ~]# chkconfig --list|grep 3:oncrond 0:off 1:off 2:on 3:on 4:on 5:on 6:offnetwork 0:off 1:off 2:on 3:on 4:on 5:on 6:offrsyslog 0:off 1:off 2:on 3:on 4:on 5:on 6:offsshd 0:off 1:off 2:on 3:on 4:on 5:on 6:off


7. timed Automatic Cleaning/var/spool/clientmqueue/Directory spam file, where inode nodes are full

In this optimization point, you can ignore the unnecessary operation on 6.4!

[root@c64 ~]# mkdir /server/scripts -p[root@c64 ~]# vi /server/scripts/spool_clean.sh#!/bin/shfind/var/spool/clientmqueue/-typef -mtime +30|xargsrm-f


Add it to the crontab scheduled task.

[root@c64 ~]# echo '*/30 * * * * /bin/sh /server/scripts/spool_clean.sh >/dev/null 2>&1'>>/var/spool/cron/root


8. Change the defaultsshService port, prohibit remote connection by the root user

[Root @ c64 ~] # Cp/etc/ssh/sshd_config/etc/ssh/sshd_config.bak [root @ c64 ~] # Vim/etc/ssh/sshd_configPort 52113 # Use ssh to connect to the default port PermitRootLogin no # the root user knows that, disable remote logon to PermitEmptyPasswords no # disable empty password logon to UseDNS no # disable DNS [root @ c64 ~] #/Etc/init. d/sshd reload # load the configuration [root @ c64 ~] # Netstat-lnt # view port information [root @ c64 ~] # Lsof-I tcp: 52113


9. Lock key file systems

[root@c64 ~]# chattr +i /etc/passwd[root@c64 ~]# chattr +i /etc/inittab[root@c64 ~]# chattr +i /etc/group[root@c64 ~]# chattr +i /etc/shadow[root@c64 ~]# chattr +i /etc/gshadow


After using the chattr command, we need to rename it for security purposes

[Root @ c64 ~] #/Bin/mv/usr/bin/chattr/usr/bin/any name


10. Adjust the file descriptor size

[Root @ localhost ~] # Ulimit-n # view the file descriptor size: 1024 [root @ localhost ~] # Echo '*-nofile 65535'>/etc/security/limits. conf


After the configuration is complete, log on again to view it.

Tip: You can also add the ulimit-SHn 65535 command to/etc/rc. local, and the command will take effect after each restart.

[root@c64 ~]# cat >>/etc/rc.local<<EOF#open filesulimit -HSn 65535#stack sizeulimit -s 65535EOF


Extension: file descriptor

The file descriptor is a non-negative integer in form. In fact, it is an index value that points to the record table for opening files for each process maintained by the kernel. When the program opens an existing file or creates a new file, the kernel returns a file descriptor to the process. In program design, some underlying programming is usually centered around the file descriptor. However, the file descriptor concept is often only applicable to operating systems such as Unix and Linux.

Traditionally, the file descriptor of the standard input is 0, the standard output is 1, and the standard error is 2. Although this habit is not a feature of the Unix kernel, many applications will not be able to use it because some shells and many applications use this habit.


11. Adjust the character set to support Chinese Characters

sed-i 's#LANG="en_US.UTF-8"#LANG="zh_CN.GB18030"#'/etc/sysconfig/i18nsource/etc/sysconfig/i18n


Extended: What is a character set?

In short, it is a set of text symbols and Their encoding. Common Character sets include:

GBK fixed-length dual-byte is not an international standard and many support systems

UTF-8 non-fixed length 1-4 bytes widely supported, MYSQL also use UTF-8


12. Remove the screen display before system and kernel version Logon

[root@c64 ~]# >/etc/redhat-release[root@c64 ~]# >/etc/issue


13. Kernel Parameter Optimization

Note: This optimization is suitable for apache, nginx, squid, and other web applications. Special Services may need to be slightly adjusted.

[Root @ c64 ~] # Vi/etc/sysctl. conf # by sun in 20131001net. ipv4.tcp _ fin_timeout = 2net. ipv4.tcp _ tw_reuse = 1net. ipv4.tcp _ tw_recycle = 1net. ipv4.tcp _ syncookies = 1net. ipv4.tcp _ keepalive_time = 600net. ipv4.ip _ local_port_range = 4000 65000net. ipv4.tcp _ max_syn_backlog = 16384net. ipv4.tcp _ max_tw_buckets = 36000net. route 4.route. gc_timeout = 100net. ipv4.tcp _ syn_retries = 1net. ipv4.tcp _ synack_retries = 1net. core. somaxconn = 16384net. core. netdev_max_backlog = 16384net. ipv4.tcp _ max_orphans = 16384 # The following parameter is an optimization for the iptables firewall. If the firewall does not have a meeting, ignore it. Net. ipv4.ip _ conntrack_max = 2520.00net. ipv4.netfilter. ip_conntrack_max = 2520.00net. ipv4.netfilter. ip_conntrack_tcp_timeout_established = 180net. ipv4.netfilter. ip_conntrack_tcp_timeout_time_wait = 120net. ipv4.netfilter. ip_conntrack_tcp_timeout_close_wait = 60net. ipv4.netfilter. ip_conntrack_tcp_timeout_fin_wafit = 120 [root @ localhost ~] # Sysctl-p # Make the configuration file take effect


Tip: The Module name in CentOS6.X is not ip_conntrack, but nf_conntrack, so in/etc/sysctl. net. ipv4.netfilter. change ip_conntrack_max to net. netfilter. nf_conntrack_max.

That is, to optimize the firewall, which is

net.ipv4.ip_conntrack_max = 25000000net.ipv4.netfilter.ip_conntrack_max = 25000000net.ipv4.netfilter.ip_conntrack_tcp_timeout_established = 180net.ipv4.netfilter.ip_conntrack_tcp_timeout_time_wait = 120net.ipv4.netfilter.ip_conntrack_tcp_timeout_close_wait = 60net.ipv4.netfilter.ip_conntrack_tcp_timeout_fin_wait = 120


On 6.4, yes

net.nf_conntrack_max = 25000000net.netfilter.nf_conntrack_max = 25000000net.netfilter.nf_conntrack_tcp_timeout_established = 180net.netfilter.nf_conntrack_tcp_timeout_time_wait = 120net.netfilter.nf_conntrack_tcp_timeout_close_wait = 60net.netfilter.nf_conntrack_tcp_timeout_fin_wait = 120


In addition, an error may be reported during the optimization process:

1. In version 5.8

error: "net.ipv4.ip_conntrack_max"is an unknown keyerror: "net.ipv4.netfilter.ip_conntrack_max"is an unknown keyerror: "net.ipv4.netfilter.ip_conntrack_tcp_timeout_established"is an unknown keyerror: "net.ipv4.netfilter.ip_conntrack_tcp_timeout_time_wait"is an unknown keyerror: "net.ipv4.netfilter.ip_conntrack_tcp_timeout_close_wait"is an unknown keyerror: "net.ipv4.netfilter.ip_conntrack_tcp_timeout_fin_wait"is an unknown key


This error may be because your firewall is not enabled or the module ip_conntrack that can be loaded is not automatically loaded. solution 2: Enable the firewall and enable the module ip_conntrack.

modprobe ip_conntrackecho "modprobe ip_conntrack">> /etc/rc.local


2. Version 6.4

error: "net.nf_conntrack_max"isan unknown keyerror: "net.netfilter.nf_conntrack_max"isan unknown keyerror: "net.netfilter.nf_conntrack_tcp_timeout_established"isan unknown keyerror: "net.netfilter.nf_conntrack_tcp_timeout_time_wait"isan unknown keyerror: "net.netfilter.nf_conntrack_tcp_timeout_close_wait"isan unknown keyerror: "net.netfilter.nf_conntrack_tcp_timeout_fin_wait"isan unknown key


This error may be because your firewall is not enabled or the module ip_conntrack that can be loaded is not automatically loaded. solution 2: Enable the firewall and enable the module ip_conntrack.

modprobe nf_conntrackecho "modprobe nf_conntrack">> /etc/rc.local


3. Version 6.4

error: "net.bridge.bridge-nf-call-ip6tables"isan unknown keyerror: "net.bridge.bridge-nf-call-iptables"isan unknown keyerror: "net.bridge.bridge-nf-call-arptables"isan unknown key


This error occurs because the system does not automatically load the loaded module. The solution is to automatically process the loaded module ip_conntrack.

modprobe bridgeecho "modprobe bridge">> /etc/rc.local

At this point, the basic optimization after Linux is installed is almost done. We need to be familiar with 13 optimization points. In the future, I will generate a one-click shell script for discussion and learning.


This article is from the "sunsky" blog, please be sure to keep this source http://sunsky.blog.51cto.com/4824967/1318607


Contact Us

The content source of this page is from Internet, which doesn't represent Alibaba Cloud's opinion; products and services mentioned on that page don't have any relationship with Alibaba Cloud. If the content of the page makes you feel confusing, please write us an email, we will handle the problem within 5 days after receiving your email.

If you find any instances of plagiarism from the community, please send an email to: info-contact@alibabacloud.com and provide relevant evidence. A staff member will contact you within 5 working days.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

  • Sales Support

    1 on 1 presale consultation

  • After-Sales Support

    24/7 Technical Support 6 Free Tickets per Quarter Faster Response

  • Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.