日常问题处理 Tips

一、kvm/OpenNebula

• onehost sync --force
  • OpenNebula host sync
• messagebus 服务必须开启，否则添加节点会失败
  • chkconfig messagebus on
  • service messagebus start
• 节点添加或监控失败调试，所在节点执行以下命令调试
  • cd /var/tmp/one/im;bash -x run_probes kvm /var/lib/one/datastores 4124 20 0 localhost
• 网卡驱动推荐virtio，提升网卡性能，默认 rtl8139
  • KVM networking performance
  • Make the Network of Your VMs Fly With the Virtio Driver

二、git/svn

• git 忽略文件.gitignore
• A guide for programming within version control.

Git 分支管理

• 列出分支
  • git branch列出本地分支
    • git branch -v
  • git branch -r列出远程分支
    • git branch -r -v
• 删除分支 「-D 表强制删除」
  • git branch -d local-branch-name删除本地分支
  • git branch -d -r remote-branch-name; git push origin :remote-branch-name删除远程分支
• 查看分支 merge 信息「-r 此处依然生效」
  • git branch --merged已 merge 分支
  • git branch --no-merged未 merge 分支
• Rebase
  • git fetch origin; git rebase origin/master

三、Linux

• grep

grep 匹配文件差异

➜  /tmp/test/ cat test1
9
10
a
b
c
➜  /tmp/test/ cat test2
6
9
10
c
x
y
z
➜  /tmp/test/ grep -F -v -f test1 test2
6
x
y
z

route

route del -net 192.168.0.0 netmask 255.255.255.0 dev eth1   # 删除指定网段路由
route add -net 192.168.0.0 netmask 255.255.255.0 dev eth1   # 添加指定网段路由
route add default gw 192.168.0.1    # 增加默认路由
route del default                   # 删除默认路由
route add -net 192.168.1.0 netmask 255.255.255.0 gw 192.168.0.1     # 增加静态路由
route del -net 192.168.1.0 netmask 255.255.255.0 gw 192.168.0.1     # 删除静态路由

ip route

ip route del 192.168.0.0/24 dev eth1    # 删除指定网段路由
ip route add 192.168.0.0/24 dev eth1    # 添加指定网段路由
ip route del via 10.2.255.254       # 删除默认路由
ip route add via 10.2.255.254       # 增加默认路由
ip route add 192.168.1.0/24 via 192.168.0.1  # 增加静态路由，192.168.0.1 为下一跳地址
ip route del 192.168.1.0/24 via 192.168.0.1  # 删除静态路由

drop caches

Writing to this will cause thekernel to drop clean caches, dentries and inodes from memory, causing thatmemory to become free.

• To free pagecache:
  • echo 1 > /proc/sys/vm/drop_caches
• To free dentries and inodes:
  • echo 2 > /proc/sys/vm/drop_caches
• To free pagecache, dentries andinodes:
  • echo 3 > /proc/sys/vm/drop_caches
• Reference: sysctl vm.txt

As this is a non-destructiveoperation and dirty objects are not freeable, the user should runsyncfirst.

swappiness

echo 0 > /proc/sys/vm/swappiness

默认值 60，对于服务器这是一个糟糕的默认值，这个值只对笔记本适用。服务器应该设置为 0。

桥接

• CentOS

# cat /etc/sysconfig/network-scripts/ifcfg-eth0
DEVICE=eth0
TYPE=Ethernet
HWADDR=xx:xx:xx:xx:xx:xx
BOOTPROTO=none
ONBOOT=yes
NM_CONTROLLED=no
BRIDGE=br0
# cat /etc/sysconfig/network-scripts/ifcfg-br0
DEVICE=br0
TYPE=Bridge
IPADDR=xx.xx.xx.xx
NETMASK=xx.xx.xx.xx
GATEWAY=xx.xx.xx.xx
ONBOOT=yes
BOOTPROTO=none
NM_CONTROLLED=no
DELAY=0
# cat /etc/sysconfig/network-scripts/ifcfg-eth1
TYPE=Ethernet
DEVICE=eth1
ONBOOT=yes
BRIDGE=br1
# cat /etc/sysconfig/network-scripts/ifcfg-br1
DEVICE=br1
TYPE=Bridge
BOOTPROTO=none
ONBOOT=yes

• Ubuntu

# cat /etc/network/interfaces
auto lo
iface lo inet loopback

auto eth0
iface eth0 inet manual

auto br0
iface br0 inet static
address xx.xx.xx.xx
netmask xx.xx.xx.xx
gateway xx.xx.xx.xx
bridge_ports eth0
bridge_stp off
bridge_fd 0
bridge_maxwait 0
dns-nameservers xx.xx.xx.xx
dns-search xxx.com

CentOS 7/RHEL 7

• 主机名配置文件
  • /etc/hostname
• ifconfig
  • yum install net-tools -y
• eth0
  • 添加选项net.ifnames=0biosdevname=0到/etc/default/grub
  • grub2-mkconfig -o /boot/grub2/grub.cfg
  • mv /etc/sysconfig/network-scripts/{ifcfg-eno16777736,ifcfg-eth0}
  • reboot

# cat /etc/default/grub
GRUB_TIMEOUT=5
GRUB_DISTRIBUTOR="$(sed 's, release .*$,,g' /etc/system-release)"
GRUB_DEFAULT=saved
GRUB_DISABLE_SUBMENU=true
GRUB_TERMINAL_OUTPUT="console"
GRUB_CMDLINE_LINUX="vconsole.keymap=us crashkernel=auto  vconsole.font=latarcyrheb-sun16 net.ifnames=0 biosdevname=0 rhgb quiet"
GRUB_DISABLE_RECOVERY="true"
# grub2-mkconfig -o /boot/grub2/grub.cfg
# mv /etc/sysconfig/network-scripts/{ifcfg-eno16777736,ifcfg-eth0}
# reboot

• diable ipv6
  • 加入ipv6.disable=1选项到/etc/default/grub配置文件中的GRUB_CMDLINE_LINUX= ...行中

查看超级块信息

• mkfs

# mkfs.ext3 -n /dev/vda1        # 非挂载情况下查看
mke2fs 1.41.12 (17-May-2010)
文件系统标签=
操作系统:Linux
块大小=1024 (log=0)
分块大小=1024 (log=0)
Stride=0 blocks, Stripe width=0 blocks
51200 inodes, 204800 blocks
10240 blocks (5.00%) reserved for the super user
第一个数据块=1
Maximum filesystem blocks=67371008
25 block groups
8192 blocks per group, 8192 fragments per group
2048 inodes per group
Superblock backups stored on blocks:
    8193, 24577, 40961, 57345, 73729

• tune2fs

# tune2fs -l /dev/vda1         # 挂载非挂载情况都可以查看
tune2fs 1.41.12 (17-May-2010)
Filesystem volume name:   /boot/
Last mounted on:          /boot
Filesystem UUID:          eaa5734f-5937-47d8-b956-e1b186f4bcd1
Filesystem magic number:  0xEF53
Filesystem revision #:    1 (dynamic)
... ...

• dumpe2fs

# dumpe2fs /dev/vda1           # 可以获取更详细的信息
dumpe2fs 1.41.12 (17-May-2010)
Filesystem volume name:   /boot/
Last mounted on:          /boot
Filesystem UUID:          eaa5734f-5937-47d8-b956-e1b186f4bcd1
Filesystem magic number:  0xEF53
Filesystem revision #:    1 (dynamic)

国内 NTP 权威网站

• www.pool.ntp.org

文件系统修复

• 强制重启 fsck 文件系统

touch /forcefsck
echo y > forcefsck
reboot

注: CentOS 5 可以使用shutdown -rF now强制修复文件系统，但是 CentOS 6 已经失效了，推荐采用上面的方式强制重启修复文件系统

• How to Check and Repair EXT3/EXT4 Filesystem on Linux

Linux OOM

• OOM Control and Notifications
• OOM Killer value always one less than set

dmidecode && MegaCli

• 查看机器型号

# dmidecode | grep "Product"

• 查看厂商

# dmidecode| grep  "Manufacturer"

• 查看序列号

# dmidecode | grep  "Serial Number"

• 查看 CPU 信息

# dmidecode | grep  "CPU"

• 查看 CPU 个数

# dmidecode | grep  "Socket Designation: CPU" |wc –l

• 查看出厂日期

# dmidecode | grep "Date"

• 查看充电状态

# MegaCli -AdpBbuCmd -GetBbuStatus -aALL |grep "Charger Status"

• 显示 BBU 状态信息

# MegaCli -AdpBbuCmd -GetBbuStatus –aALL

• 显示 BBU 容量信息

# MegaCli -AdpBbuCmd -GetBbuCapacityInfo –aALL

• 显示 BBU 设计参数

# MegaCli -AdpBbuCmd -GetBbuDesignInfo –aALL

• 显示当前 BBU 属性

# MegaCli -AdpBbuCmd -GetBbuProperties –aALL

• 查看充电进度百分比

# MegaCli -AdpBbuCmd -GetBbuStatus -aALL |grep "Relative State of Charge"

• 查询 Raid 阵列数

# MegaCli -cfgdsply -aALL |grep "Number of DISK GROUPS:"

• 显示 Raid 卡型号，Raid 设置，Disk 相关信息

# MegaCli -cfgdsply –aALL

• 显示所有物理信息

# MegaCli -PDList -aALL

• 显示所有逻辑磁盘组信息

# MegaCli -LDInfo -LALL –aAll

• 查看物理磁盘重建进度(重要)

# MegaCli -PDRbld -ShowProg -PhysDrv [1:5] -a0

• 查看适配器个数

# MegaCli –adpCount

• 查看适配器时间

# MegaCli -AdpGetTime –aALL

• 显示所有适配器信息

# MegaCli -AdpAllInfo –aAll

• 查看 Cache 策略设置

# MegaCli -cfgdsply -aALL |grep Polic

更多参考:DELL磁盘阵列控制卡（RAID卡）MegaCli常用管理命令汇总

root ulimit

/etc/init.d/ 下 root 用户启动的服务还和/etc/sysconfig/init相关，默认 nofile1024「但是该文件不建议修改，可以考虑使用非 root 用户管理服务或者针对相应服务作修改」

RubyGems

$ gem sources --remove https://rubygems.org/
$ gem sources -a https://ruby.taobao.org/
$ gem sources -l
*** CURRENT SOURCES ***

https://ruby.taobao.org
# 请确保只有 ruby.taobao.org
$ gem install rails

ssh

• 通过 ssh 私钥获取公钥

sh-keygen -y -f .ssh/id_rsa

• ~/.ssh/config

➜  ~  cat ~/.ssh/config
HOST *
    Port 22
    User root
    controlmaster auto
    controlPath ~/.ssh/master-%r@%h:%p
    ServerAliveInterval 30
    StrictHostKeyChecking no

SysVinit to Systemd Cheatsheet

• SysVinit to Systemd Cheatsheet

四、MySQL

mysqldump

• 全备

mysqldump --hex-blob --single-transaction --alldatabase --master-data=1 > all.sql

Dump binary strings (BINARY, VARBINARY, BLOB) in hexadecimal format (for example, ′abc′ becomes 0x616263). The affected data types are BINARY, VARBINARY, the BLOB types, and BIT.

• mysqldump --ignore-database

mysqldump --databases `mysql -uroot --skip-column-names \
    -e "SELECT GROUP_CONCAT(schema_name SEPARATOR ' ') \
    FROM information_schema.schemata WHERE schema_name \
    NOT IN ('mysql','performance_schema','information_schema');"` \
    > backup.sql

OR

echo '[mysqldump]' > mydump.cnf
mysql -NBe "select concat('ignore-table=', table_schema, '.', table_name) \
    from information_schema.tables \
    where table_schema in ('mysql', 'personnel', 'buildings')" \
    >> mydump.cnf

Now the options file looks like this:

[mysqldump]
ignore-table=mysql.db
ignore-table=mysql.host
ignore-table=mysql.user
[...]

mysqldump --defaults-file=./mydump.cnf  -u $DBUSER -p$DBPWD --all-databases

• 只备份表结构 [--no-data] 选项

mysqldump --no-data --databases ... ...

mysqlbinlog

• row 格式查看

mysqlbinlog -v -v --base64-output=DECODE-ROWS binlog文件名

从库跳过错误日志

mysql> STOP SLAVE;
mysql> SET GLOBAL SQL_SLAVE_SKIP_COUNTER = 1; # 1 表示跳过 1 个 events
mysql> START SLAVE;

ROW_FORMAT

• 查看当前数据库下的所有表的 ROW_FORMAT

mysql> SELECT `table_name`, `row_format` FROM `information_schema`.`tables` WHERE `table_schema`=DATABASE();

• 查看指定表 ROW_FORMAT

mysql> SHOW TABLE STATUS LIKE 'table_name'\G

MySQL 读取配置文件顺序

$ mysql --help | grep '/etc/my.cnf' # 相同选项以最后一个配置文件选项为主
/etc/my.cnf /etc/mysql/my.cnf /usr/local/etc/my.cnf ~/.my.cnf

五、Nginx

• 域名子目录访问跳转到其它网站

server {
... ...
           location ^~ /test/ {
               rewrite ^/test/(.*) /$1 break;
               proxy_pass http://www.test.com;
           }
... ...
}

• 404 页面 301 重定向到主页

server {
... ...
           location / {
               error_page 404 =  @foobar;
           }

           location @foobar {
               rewrite  .*  / permanent;
           }
... ...
}

• Nginx 配置 ip 直接访问的默认站点

server {
    ... ...
    listen 80 default;
    ... ...
}

• uwsgi

server {
    listen   80;
    server_name www.test.com;

    location / {
        include uwsgi_params;
        uwsgi_pass 127.0.0.1:8000;
    }

    location /static {
        alias /data/www/project/app/static;
    }
    ... ...
}

六、Golang

go get

• go get
  • 获取第三方包
• go get -u all
  • update the third package
• go get

go 命令标准教程

• 标准命令详解

七、Other

jekyll

• {{}} 转义

jekyll 在解析 Markdown 转义的时候会把 {{}} 识别为模板变量，所以需要通过转义解决这个问题：

{% raw %}Hello, my name is {{name}}.{% endraw %}