mdadm software raid

huataihuang · Aug 16, 2023 · b99f996 · b99f996
1 parent d96e0e1
commit b99f996
Show file tree

Hide file tree

Showing 85 changed files with 1,090 additions and 17 deletions.
diff --git a/source/_static/linux/redhat_linux/systemd/systemd_mount.png b/source/_static/linux/redhat_linux/systemd/systemd_mount.png
diff --git a/source/apple/one/apple_news_in_cn.rst b/source/apple/one/apple_news_in_cn.rst
@@ -8,7 +8,13 @@
 
    通过 :ref:`apple_one` 订阅，可以同时获得苹果的6个服务，如果购买 Premier plan ，每月 29.95 美刀，可以家庭共享并且获得 2T icloud 存储。通过 Apple TV 和 Apple News ，可以打开英语世界的大门...
 
-   美区账号无法直接使用国内的信用卡，即使是双币信用卡(假的Visa/Master?)也无法在Google/Apple上直接订阅，变通的曲径是采用 `Apple Gift Card - App Store, iTunes, iPhone, iPad, AirPods, MacBook, accessories and more (Email Delivery) <https://www.amazon.com/gp/product/B08F7GTP2R/ref=ppx_yo_dt_b_asin_title_o00?ie=UTF8&psc=1>`_ 向美区账号充值
+   美区账号无法直接使用国内的信用卡，即使是双币信用卡(假的Visa/Master?)也无法在Google/Apple上直接订阅，变通的曲径有两种方法:
+
+   - 采用 `Apple Gift Card - App Store, iTunes, iPhone, iPad, AirPods, MacBook, accessories and more (Email Delivery) <https://www.amazon.com/gp/product/B08F7GTP2R/ref=ppx_yo_dt_b_asin_title_o00?ie=UTF8&psc=1>`_ 向美区账号充值(目前我采用这种方法)
+   - `美区 Apple ID 绑定 美区 Paypal <https://blog.shuziyimin.org/171>`_ 这个方法可能更好，但是有一个非常难以跨越的门槛(我没有成功): 需要有一个美国手机号(也可以是虚拟号码)
+
+     - `Google Voice <https://voice.google.com/>`_ 虚拟号码的依赖条件是首先有一个美国移动电话号码，晕倒...我就是要一个虚拟号码来替代美国手机号码啊...
+     - `TextNow <https://www.textnow.com/>`_ 虚拟号码似乎屏蔽了云计算服务的虚拟机IP导致无法注册...
 
 准备和加载Apple News
 =====================

diff --git a/source/clang/upgrade_gcc_on_centos7.rst b/source/clang/upgrade_gcc_on_centos7.rst
@@ -18,6 +18,10 @@
 .. literalinclude:: upgrade_gcc_on_centos7/build_gcc
    :caption: 编译gcc
 
+.. warning::
+
+   编译GCC是非常大型的编译，如果没有使用 :ref:`parallel_make` 则编译过程非常漫长。对于多处理器核心建议使用 :ref:`parallel_make` 。不过，也很郁闷的是，这个并行编译可能还受限于gcc版本，我在aliOS 7.2环境使用发行版的gcc 4.8.5，开启 :ref:`parallel_make` 居然无法完成 ``gcc-10.5.0`` 编译，取消了并行编译才能正确完成编译gcc。
+
 参考
 ======
 

diff --git a/source/clang/upgrade_gcc_on_centos7/prepare_build_gcc b/source/clang/upgrade_gcc_on_centos7/prepare_build_gcc
@@ -1 +1 @@
-sudo yum -y install bzip2 wget gcc gcc-c++ gmp-devel mpfr-devel libmpc-devel make
+sudo yum -y install bzip2 wget gcc gcc-c++ gmp-devel mpfr-devel libmpc-devel make texinfo
diff --git a/source/flask/flask_startup.rst b/source/flask/flask_startup.rst
@@ -52,7 +52,11 @@ Flask 使用的 HTML 渲染模版 `Jinja2 template <http://quintagroup.com/cms/p
    :caption: 明确编写 ``escape()`` 防止恶意注入JS
    :emphasize-lines: 2,8
 
-此时用户在浏览器中输入 http:://127.0.0.1:5000/<script>alert("bad")</script> 这样的注入，就会直接被拒绝渲染，页面提示::
+.. note::
+
+   程序中 ``return f"Hello, {escape(name)}!"`` 中有一个 ``f`` 表示 ``f-string`` ，是Python 3.6以上版本共鞥，用于格式化字符串。也就是最后返回的值( ``escape()`` 处理后的值进行转义 )，最后返回的是 ``User xxx`` 
+
+此时用户在浏览器中输入 http:://127.0.0.1:5001/<script>alert("bad")</script> 这样的注入，就会直接被拒绝渲染，页面提示::
 
    Not Found
    The requested URL was not found on the server. If you entered the URL manually please check your spelling and try again.
@@ -61,12 +65,34 @@ Flask 使用的 HTML 渲染模版 `Jinja2 template <http://quintagroup.com/cms/p
 
    注意， `Jinja2 template <http://quintagroup.com/cms/python/jinja2>`_ 已经内嵌了自动防范的功能，上述代码片段不使用 ``escape()`` 也是有同样效果的
 
-所以用户只能输入正常的名字 如 http:://127.0.0.1:5000/huatai
+所以用户只能输入正常的名字 如 http:://127.0.0.1:5001/huatai
 
 此时页面才能正常渲染::
 
    Hello,huatai!
 
+路由(Routing)
+================
+
+Web应用会使用一些有意义的URLs让用户访问以及调用不同的函数返回页面，这种方式称为 ``route`` （路由)
+
+.. literalinclude:: flask_startup/hello_route.py
+   :language: python
+   :caption: 路由访问
+
+不同规则
+==========
+
+以下代码可以从url中获取需要的内容，根据不同路径以及关系分别返回 ``username / post_id / subpath`` ，你可以分别试试::
+
+   http://http://127.0.0.1:5001/user/huatai
+   http://http://127.0.0.1:5001/post/1
+   http://http://127.0.0.1:5001/path/the_large_path/small_path
+
+.. literalinclude:: flask_startup/variable_rules.py
+   :language: python
+   :caption: 根据用户访问路径来返回不同类型的数据
+
 参考
 =======
 

diff --git a/source/flask/flask_startup/hello_route.py b/source/flask/flask_startup/hello_route.py
@@ -0,0 +1,11 @@
+from flask import Flask
+
+app = Flask(__name__)
+
+@app.route('/')
+def index():
+    return 'Index Page'
+
+@app.route("/hello")
+def hello_world():
+    return "Hello, World"
diff --git a/source/flask/flask_startup/variable_rules.py b/source/flask/flask_startup/variable_rules.py
@@ -0,0 +1,19 @@
+from flask import Flask
+from markupsafe import escape
+
+app = Flask(__name__)
+
+@app.route('/user/<username>')
+def show_user_profile(username):
+    # show the user profile for that user
+    return f'User {escape(username)}'
+
+@app.route('/post/<int:post_id>')
+def show_post(post_id):
+    # show the post with the given id, the id is an interger
+    return f'Post {post_id}'
+
+@app.route('/path/<path:subpath>')
+def show_subpath(subpath):
+    # show the subpath after /path/
+    return f'Subpath {escape(subpath)}'
diff --git a/source/flask/intro_flask.rst b/source/flask/intro_flask.rst
@@ -15,7 +15,7 @@ Flask依赖以下开源项目:
 - `MarkupSafe <https://palletsprojects.com/p/markupsafe/>`_ 随Jinja提供，渲染模版是避免注入攻击
 - `ItsDangerous <https://palletsprojects.com/p/itsdangerous/>`_ 对数据进行安全签名确保完整性，用于保护flask的会话cookie
 - `click command line interfaces <https://click.palletsprojects.com/>`_ 编写命令行应用程序的框架，提供了flask命令以及允许添加定制的管理命令
-- `Blinker <https://blinker.readthedocs.io/>` 实现 ``Signals`` 功能(轻量级订阅通知)
+- `Blinker <https://blinker.readthedocs.io/>`_ 实现 ``Signals`` 功能(轻量级订阅通知)
 
 Flask 并不原生支持数据库、Web表单、用户认证等高级功能，这些功能以及大多数Web程序所需的关键服务都以扩展的方式集成到核心包:
 

diff --git a/source/gluster/best_practices_for_gluster/gluster_underlay_filesystem.rst b/source/gluster/best_practices_for_gluster/gluster_underlay_filesystem.rst
@@ -32,4 +32,4 @@ GlusterFS底层需要采用操作系统提供的文件系统，有以下推荐
 实践
 =======
 
-- :ref:`deploy_centos7_gluster11_lvm_mdadm_raid6`
+- :ref:`deploy_centos7_gluster11_lvm_mdadm_raid10`
diff --git a/source/gluster/deploy/centos/centos_gluster_init.rst b/source/gluster/deploy/centos/centos_gluster_init.rst
@@ -67,6 +67,12 @@ CentOS目前依然是生产环境中常用的操作系统，虽然由于 :ref:`r
    :language: bash
    :caption: 通过 :ref:`pssh` 批量完成分区、格式化和挂载
 
+- 完成后在服务器上检查 ``df -h`` 可以看到12块已经格式化好 :ref:`xfs` 并挂载好的磁盘:
+
+.. literalinclude:: centos_gluster_init/parted_xfs_df
+   :caption: 分区和格式化、挂载好的12块磁盘的文件系统
+   :emphasize-lines: 9-20
+
 ``wrong fs type, bad option, bad superblock``
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 

diff --git a/source/gluster/deploy/centos/centos_gluster_init/parted_xfs.sh b/source/gluster/deploy/centos/centos_gluster_init/parted_xfs.sh
@@ -1,6 +1,10 @@
 for i in {0..11};do
     if [ ! -d /data/brick${i} ];then mkdir -p /data/brick${i};fi
     parted -s -a optimal /dev/nvme${i}n1 mklabel gpt
+    # 如果随机遇到以下报错，显示磁盘设备busy无法分区，则添加sleep 1避免上一个parted命令还没处理完就发起下一个parted
+    # Error: Error informing the kernel about modifications to partition /dev/nvme1n1p1 -- Device or resource busy.  This means Linux won't know about any changes you made to /dev/nvme1n1p1 until you reboot -- so you shouldn't mount it or use it in any way before rebooting.
+    # Error: Failed to add partition 1 (Device or resource busy)
+    sleep 1
     parted -s -a optimal /dev/nvme${i}n1 mkpart primary xfs 0% 100%
     parted -s -a optimal /dev/nvme${i}n1 name 1 gluster_brick${i}
     sleep 1

diff --git a/source/gluster/deploy/centos/centos_gluster_init/parted_xfs_df b/source/gluster/deploy/centos/centos_gluster_init/parted_xfs_df
@@ -0,0 +1,20 @@
+Filesystem       Size  Used Avail Use% Mounted on
+devtmpfs          63G     0   63G   0% /dev
+tmpfs             63G   66M   63G   1% /dev/shm
+tmpfs             63G  2.2M   63G   1% /run
+tmpfs             63G     0   63G   0% /sys/fs/cgroup
+/dev/sda3         49G   21G   27G  44% /
+/dev/sda5        167G  1.4G  157G   1% /home
+/dev/sda2        976M  181M  728M  20% /boot
+/dev/nvme0n1p1   3.5T  3.7G  3.5T   1% /data/brick0
+/dev/nvme1n1p1   3.5T  3.7G  3.5T   1% /data/brick1
+/dev/nvme2n1p1   3.5T  3.7G  3.5T   1% /data/brick2
+/dev/nvme3n1p1   3.5T  3.7G  3.5T   1% /data/brick3
+/dev/nvme4n1p1   3.5T  3.7G  3.5T   1% /data/brick4
+/dev/nvme5n1p1   3.5T  3.7G  3.5T   1% /data/brick5
+/dev/nvme6n1p1   3.5T  3.7G  3.5T   1% /data/brick6
+/dev/nvme7n1p1   3.5T  3.7G  3.5T   1% /data/brick7
+/dev/nvme8n1p1   3.5T  3.7G  3.5T   1% /data/brick8
+/dev/nvme9n1p1   3.5T  3.7G  3.5T   1% /data/brick9
+/dev/nvme10n1p1  3.5T  3.7G  3.5T   1% /data/brick10
+/dev/nvme11n1p1  3.5T  3.7G  3.5T   1% /data/brick11
diff --git a/...loy_centos7_gluster11_lvm_mdadm_raid6.rst → ...oy_centos7_gluster11_lvm_mdadm_raid10.rst b/...loy_centos7_gluster11_lvm_mdadm_raid6.rst → ...oy_centos7_gluster11_lvm_mdadm_raid10.rst
@@ -1,21 +1,32 @@
-.. _deploy_centos7_gluster11_lvm_mdadm_raid6:
+.. _deploy_centos7_gluster11_lvm_mdadm_raid10:
 
-========================================
-在软RAID6+LVM上CentOS 7 部署Gluster 11
-========================================
+=============================================
+在软RAID10 + LVM上CentOS 7 部署Gluster 11
+=============================================
 
 .. note::
 
    根据 :ref:`deploy_centos7_gluster11` 迭代改进部署方案
 
-**待续**
+在 :ref:`deploy_centos7_gluster11` 我采用了一个比较ugly的部署方案，将物理服务器上的12个 ``brick`` (磁盘)分别用于同一个volume，但是由于强制顺序安排，可以让数据分布到不同服务器上。但是，实际上这种方案限制了集群的扩容和缩容，例如在 :ref:`add_centos7_gluster11_server` 就会看到缺陷。对于GlusterFS这种精简架构的分布式存储， :ref:`think_best_practices_for_gluster` ，改进为底层采用 :ref:`linux_software_raid` 来统一存储磁盘，实现一个超大规模的磁盘，然后借助 :ref:`linux_lvm` 来实现灵活的卷划分和管理。
 
 准备工作
 ===========
 
 - :ref:`build_glusterfs_11_for_centos_7`
 - :ref:`gluster11_rpm_createrepo`
 
+磁盘存储池构建
+================
+
+- :ref:`mdadm_raid10` 是因为本省存储磁盘非常充裕，所以我就没有采用更为节约磁盘(且高数据安全)的RAID6
+
+.. warning::
+
+   由于项目方案调整，我现在采用了直接在 :ref:`mdadm_raid10` 构建 :ref:`xfs` ，快速完成部署。
+
+   以下部分请暂时忽略，等后续再有实践机会重新开始...
+
 安装和启动服务
 ===============
 

diff --git a/source/gluster/deploy/centos/index.rst b/source/gluster/deploy/centos/index.rst
@@ -12,7 +12,7 @@ CentOS平台GlusterFS部署
    download_gluster_rpm_createrepo.rst
    gluster11_rpm_createrepo.rst
    deploy_centos7_gluster11.rst
-   deploy_centos7_gluster11_lvm_mdadm_raid6.rst
+   deploy_centos7_gluster11_lvm_mdadm_raid10.rst
    add_centos7_gluster11_server.rst
 
 .. only::  subproject and html

diff --git a/source/kernel/process/utils/ps.rst b/source/kernel/process/utils/ps.rst
@@ -15,10 +15,41 @@ Linux/Unix 通常采用 System V ( ``ps -elf`` ) 或 BSD ( ``ps alx`` ) 风格 `
 
    似乎 ``ps -elf`` 更为适合(输出字段中有 ``C`` 表明CPU使用率)
 
-后续再学习
+- 检查进程启动时间 ``lstart`` :
 
+.. literalinclude:: ps/ps_lstart
+   :caption: 检查进程启动时间，例如 ``qemu-system-x86``
+
+输出显示案例类似如下:
+
+.. literalinclude:: ps/ps_lstart_output
+   :caption: 检查进程启动时间，例如 ``qemu-system-x86`` 可以看到详细的启动时间
+
+``ps`` 检查线程
+=================
+
+``ps`` 命令的 ``-T`` 参数表示输出线程， ``-p`` 可以指定进程，结合上文的输出字段，我们可以构建一个检查某个进程所有线程的CPU使用率以及运行在哪些cpu core上，以便进一步排查进程异常。
+
+举例，进程 ``qemu-system-x86`` 的 ``pid`` 是 ``7354`` ，当前 ``top`` 可以看到使用的CPU百分比大约是 10+% :
+
+.. literalinclude:: ps/top_qemu
+   :caption: 检查 ``qemu-system-x86`` 进程的线程负载
+   :emphasize-lines: 8
+
+- 现在来解析这个进程的线程:
+
+.. literalinclude:: ps/ps_tid_cpu
+   :caption: 检查进程的所有线程使用的cpu资源以及调度的cpu core
+
+输出信息:
+
+.. literalinclude:: ps/ps_tid_cpu_output
+   :caption: 检查进程的所有线程使用的cpu资源以及调度的cpu core
+
+这里可以看到4个kvm线程分别消耗了大约 2.5% 的CPU资源
 
 参考
 ======
 
 - `About the output fields of the ps command in Unix <https://kb.iu.edu/d/afnv>`_ 非常清晰的 ``ps -o`` 参数字段快速查询，建议参考
+- `How to get the start time of a long-running Linux process? <https://stackoverflow.com/questions/5731234/how-to-get-the-start-time-of-a-long-running-linux-process>`_
diff --git a/source/kernel/process/utils/ps/ps_lstart b/source/kernel/process/utils/ps/ps_lstart
@@ -0,0 +1 @@
+ps -eo pid,lstart,cmd | grep qemu-system-x86_64
diff --git a/source/kernel/process/utils/ps/ps_lstart_output b/source/kernel/process/utils/ps/ps_lstart_output
@@ -0,0 +1,3 @@
+   7300 Sun Aug  6 11:10:56 2023 /usr/bin/qemu-system-x86_64 -name guest=z-b-data-3,...
+   7354 Sun Aug  6 11:11:03 2023 /usr/bin/qemu-system-x86_64 -name guest=z-b-data-1,...
+   7405 Sun Aug  6 11:11:09 2023 /usr/bin/qemu-system-x86_64 -name guest=z-b-data-2,...
diff --git a/source/kernel/process/utils/ps/ps_tid_cpu b/source/kernel/process/utils/ps/ps_tid_cpu
@@ -0,0 +1 @@
+ps -T -o pid,tid,c,pcpu,comm -p 7354
diff --git a/source/kernel/process/utils/ps/ps_tid_cpu_output b/source/kernel/process/utils/ps/ps_tid_cpu_output
@@ -0,0 +1,10 @@
+    PID     TID  C %CPU COMMAND
+   7354    7354  0  0.2 qemu-system-x86
+   7354    7358  0  0.0 qemu-system-x86
+   7354    7361  0  0.0 IO mon_iothread
+   7354    7363  2  2.1 CPU 0/KVM
+   7354    7364  2  2.6 CPU 1/KVM
+   7354    7365  2  2.5 CPU 2/KVM
+   7354    7366  2  2.7 CPU 3/KVM
+   7354    7470  0  0.0 worker
+   7354  210749  0  0.0 worker
diff --git a/source/kernel/process/utils/ps/top_qemu b/source/kernel/process/utils/ps/top_qemu
@@ -0,0 +1,11 @@
+top - 16:30:30 up 4 days,  5:28, 11 users,  load average: 0.72, 0.55, 0.40
+Tasks: 630 total,   1 running, 629 sleeping,   0 stopped,   0 zombie
+%Cpu(s):  0.6 us,  0.3 sy,  0.0 ni, 99.1 id,  0.0 wa,  0.0 hi,  0.0 si,  0.0 st
+MiB Mem : 386813.0 total, 327264.1 free,  51822.8 used,   7726.1 buff/cache
+MiB Swap:      0.0 total,      0.0 free,      0.0 used. 332638.5 avail Mem
+
+    PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND
+   7354 libvirt+  20   0   16.8g  16.0g  19768 S  10.2   4.2 635:51.79 qemu-system-x86
+   7300 libvirt+  20   0   16.7g  16.0g  19668 S   6.6   4.2 459:39.34 qemu-system-x86
+   7405 libvirt+  20   0   16.8g  16.0g  19732 S   6.6   4.2 442:22.15 qemu-system-x86
+   2014 prometh+  20   0  726984  19972  12036 S   4.3   0.0  50:29.36 node_exporter
diff --git a/source/kernel/startup/index.rst b/source/kernel/startup/index.rst
@@ -10,6 +10,7 @@ Kernel起步
    howto_learn_kernel.rst
    prepare_kernel_dev.rst
    kernel_overview.rst
+   sysctl.rst
 
 .. only::  subproject and html
 

diff --git a/source/kernel/startup/sysctl.rst b/source/kernel/startup/sysctl.rst
@@ -0,0 +1,40 @@
+.. _sysctl:
+
+=================
+sysctl
+=================
+
+``sysctl`` 命令工具可以动态修改内核参数
+
+快速起步
+==========
+
+- 列出所有系统内核变量(参数值):
+
+.. literalinclude:: sysctl/sysctl_a
+   :caption: 列出所有内核变量及值
+
+- 读取变量(这里案例是读取 ``kernel.version`` ):
+
+.. literalinclude:: sysctl/sysctl_read
+   :caption: 读取 ``kernel.version`` 内核参数
+
+- 临时修改内核变量:
+
+.. literalinclude:: sysctl/sysctl_change
+   :caption: 临时修改内核参数
+
+- 修改内核变量持久化(也就是写入配置文件，重启依然生效):
+
+.. literalinclude:: sysctl/sysctl_change_permanently
+   :caption: 永久修改内核参数
+
+.. note::
+
+   内核参数除了配置在 ``/etc/sysctl.conf`` 配置文件，也可以将大配置文件拆解成多个存储在 ``/etc/sysctl.d`` 目录下的各个配置文件
+
+
+参考
+======
+
+- `Product Documentation > Red Hat Enterprise Linux > 7 > Kernel Administration Guide > Chapter 2. Working with sysctl and kernel tunables <https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html/kernel_administration_guide/working_with_sysctl_and_kernel_tunables>`_
diff --git a/source/kernel/startup/sysctl/sysctl_a b/source/kernel/startup/sysctl/sysctl_a
@@ -0,0 +1 @@
+sysctl -a
diff --git a/source/kernel/startup/sysctl/sysctl_change b/source/kernel/startup/sysctl/sysctl_change
@@ -0,0 +1 @@
+sysctl <tunable class>.<tunable>=<value>
diff --git a/source/kernel/startup/sysctl/sysctl_change_permanently b/source/kernel/startup/sysctl/sysctl_change_permanently
@@ -0,0 +1 @@
+sysctl -w <tunable class>.<tunable>=<value> >> /etc/sysctl.conf
diff --git a/source/kernel/startup/sysctl/sysctl_read b/source/kernel/startup/sysctl/sysctl_read
@@ -0,0 +1 @@
+sysctl kernel.version
diff --git a/source/kubernetes/network/index.rst b/source/kubernetes/network/index.rst
@@ -14,6 +14,7 @@ Kubernetes只支持基于Container Network Interface(CNI)的网络，需要通
    k8s_network_infa.rst
    k8s_loadbalancer_ingress.rst
    dynamic_dns_loadbalancing_without_cloud_provider.rst
+   k8s_hosts_file_for_pods.rst
    ingress/index
    flannel/index
    cilium/index

diff --git a/source/kubernetes/network/k8s_hosts_file_for_pods.rst b/source/kubernetes/network/k8s_hosts_file_for_pods.rst
@@ -0,0 +1,43 @@
+.. _k8s_hosts_file_for_pods:
+
+================================
+Kubernetes配置Pods中的hosts文件
+================================
+
+在故障排查中(也可能测试环境)，需要为 pod 中注入 ``/etc/hosts`` 配置，以便能够绕过DNS解析来直接访问目标服务器。在Kubernetes中， ``PodSpec`` 段落提供了 ``HostAliases`` 字段完成配置:
+
+.. literalinclude:: k8s_hosts_file_for_pods/hostalias.yaml
+   :language: yaml
+   :caption: 为pod添加hosts
+
+- 测试:
+
+.. literalinclude:: k8s_hosts_file_for_pods/kubectl_hostalias
+   :caption: 创建测试pods
+
+- 检查:
+
+.. literalinclude:: k8s_hosts_file_for_pods/kubectl_get_hostalias
+   :caption: 查看创建的测试pods
+
+输出可以看到运行的pod:
+
+.. literalinclude:: k8s_hosts_file_for_pods/kubectl_get_hostalias_output
+   :caption: 查看创建的测试pods
+
+
+- 查看日志:
+
+.. literalinclude:: k8s_hosts_file_for_pods/kubectl_logs_hostalias
+   :caption: 查看创建的测试pods的日志
+
+- 日志结果(就是 ``cat /etc/hosts`` 的输出内容验证):
+
+.. literalinclude:: k8s_hosts_file_for_pods/kubectl_logs_hostalias_output
+   :caption: 查看创建的测试pods的日志内容就是 hosts 内容
+
+
+参考
+=======
+
+- `Adding entries to Pod /etc/hosts with HostAliases <https://kubernetes.io/docs/tasks/network/customize-hosts-file-for-pods/>`_