mirror of
https://github.com/LCTT/TranslateProject.git
synced 2025-03-24 02:20:09 +08:00
MjSeven
GitHub
commit
31fc1ef838
@ -1,53 +0,0 @@
|
||||
Translating by qhwdw
|
||||
3 ways robotics affects the CIO role
|
||||
======
|
||||
|
||||
As 2017 comes to a close, many CIOs are solidifying their goals for 2018. Perhaps yours involve robotic process automation (RPA.) For years, RPA has been a distant concept for many companies. But as organizations are forced to become ever more nimble and efficient, the potential benefits of RPA bear examining.
|
||||
|
||||
According to a recent[ survey][1] by Redwood Software and Sapio Research, IT decision makers believe that 59 percent of business processes can be automated in the next five years, creating new speed and efficiency while relieving their human counterparts of repetitive manual workloads. However, 20 percent of companies with more than 1000 employees currently have no RPA strategy in place.
|
||||
|
||||
For CIOs, RPA has implications for your role in the business and for your team. Here are three ways that the role of the CIO and other IT decision makers can change as RPA gains prominence in 2018:
|
||||
|
||||
### Added opportunity to be strategic change agents
|
||||
|
||||
As the pressure grows to do more with less, internal processes matter greatly. In every enterprise, employees across departments are performing critical - yet mundane - tasks every single day. These tasks may be boring and repetitive, but they must be performed quickly, and often with no room for error.
|
||||
|
||||
**[ For advice on maximizing your automation strategy's ROI, see our related article,[How to improve ROI on automation: 4 tips][2]. ]**
|
||||
|
||||
From back-office operations in finance to procurement, supply chain, accounting, customer service, and human resources, nearly every position within an organization is plagued with at least some monotonous tasks. For CIOs, this opens up an opportunity to unite the business with IT and spearhead strategic change with RPA.
|
||||
|
||||
Having evolved far beyond screen-scraping technology of the past, robots are now customizable, plug-and-play solutions that can be built to an organization's specific needs. With such a process-centric approach, companies can automate not only tasks previously executed by humans, but also application and system-specific tasks, such as ERP and other enterprise applications.
|
||||
|
||||
Enabling a greater level of automation for end-to-end processes is where the value lies. CIOs will be on the front line of this opportunity.
|
||||
|
||||
### Renewed focus on people and training
|
||||
|
||||
Technology shifts can be unnerving to employees, especially when these changes involve automating substantial portions of their daily duties. The CIO should articulate how RPA will change roles and responsibilities for the better, and fuel data-driven, strategic decisions that will ultimately impact the bottom line.
|
||||
|
||||
When implementing RPA, it's important to convey that humans will always be critical to the success of the organization, and that success requires the right balance of technology and human skills.
|
||||
|
||||
CIOs should also analyze workflow and implement better processes that go beyond mimicking end-user specific tasks. Through end-to-end process automation, CIOs can enable employees to shine.
|
||||
|
||||
Because it will be important to upskill and retrain employees throughout the automation process, CIOs must be prepared to collaborate with the C-suite to determine training programs that help employees navigate the change with confidence.
|
||||
|
||||
### Demand for long-term thinking
|
||||
|
||||
To succeed with robotic process automation, brands must take a long-term approach. This will require a scalable solution, which in turn will benefit the entire business model, including customers. When Amazon introduced faster delivery options for Prime customers, for example, it didn't just retool the entire order fulfillment process in its warehouses; it automated its online customer experience to make it simpler, faster, and easier than ever for consumers to place orders.
|
||||
|
||||
In the coming year, CIOs can approach technology in the same way, architecting holistic solutions to change the way an organization operates. Reducing headcount will net only so much in bottom-line results, but process automation allows CIOs to think bigger through optimization and empowerment. This approach gives CIOs the opportunity to build credibility for the long haul, for themselves and for RPA. This in turn will enhance the CIO's role as a navigator and contributor to the organization's overall success.
|
||||
|
||||
For CIOs, taking a long-term, strategic approach to RPA success takes time and hard work. Nevertheless, CIOs who commit the time to create a strategy that balances manpower and technology will deliver value now and in the future.
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://enterprisersproject.com/article/2017/11/3-ways-robotics-affects-cio-role
|
||||
|
||||
作者：[Dennis Walsh][a]
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://enterprisersproject.com/user/dennis-walsh
|
||||
[1]:https://www.redwood.com/press-releases/it-decision-makers-speak-59-of-business-processes-could-be-automated-by-2022/
|
||||
[2]:https://enterprisersproject.com/article/2017/11/how-improve-roi-automation-4-tips?sc_cid=70160000000h0aXAAQ
|
||||
@ -1,112 +0,0 @@
|
||||
Translating by qhwdw
|
||||
Dynamic Linux Routing with Quagga
|
||||
============================================================
|
||||
|
||||
|
||||
|
||||
Learn how to use the Quagga suite of routing protocols to manage dynamic routing.[Creative Commons Attribution][1][Wikimedia Commons: Martin Grandjean][2]
|
||||
|
||||
So far in this series, we have learned the intricacies of IPv4 addressing in [Linux LAN Routing for Beginners: Part 1][4] and how to create static routes manually in [Linux LAN Routing for Beginners: Part 2][5].
|
||||
|
||||
Now we're going to use [Quagga][6] to manage dynamic routing for us, just set it and forget it. Quagga is a suite of routing protocols: OSPFv2, OSPFv3, RIP v1 and v2, RIPng, and BGP-4, which are all managed by the zebra daemon.
|
||||
|
||||
OSPF means Open Shortest Path First. OSPF is an interior gateway protocol (IGP); it is for LANs and LANs connected over the Internet. Every OSPF router in your network contains the topology for the whole network, and calculates the best paths through the network. OSPF automatically multicasts any network changes that it detects. You can divide up your network into areas to keep routing tables manageable; the routers in each area only need to know the next hop out of their areas rather than the entire routing table for your network.
|
||||
|
||||
RIP, Routing Information Protocol, is an older protocol. RIP routers periodically multicast their entire routing tables to the network, rather than just the changes as OSPF does. RIP measure routes by hops, and sees any destination over 15 hops as unreachable. RIP is simple to set up, but OSPF is a better choice for speed, efficiency, and scalability.
|
||||
|
||||
BGP-4 is the Border Gateway Protocol version 4\. This is an exterior gateway protocol (EGP) for routing Internet traffic. You won't use BGP unless you are an Internet service provider.
|
||||
|
||||
### Preparing for OSPF
|
||||
|
||||
In our little KVM test lab, there are two virtual machines representing two different networks, and one VM acting as the router. Create two networks: net1 is 192.168.110.0/24 and net2 is 192.168.120.0/24\. It's all right to enable DHCP because you are going to go into your three virtual machines and give each of them static addresses. Host 1 is on net1, Host 2 is on net2, and Router is on both networks. Give Host 1 a gateway of 192.168.110.126, and Host 2 gets 192.168.120.136.
|
||||
|
||||
* Host 1: 192.168.110.125
|
||||
|
||||
* Host 2: 192.168.120.135
|
||||
|
||||
* Router: 192.168.110.126 and 192.168.120.136
|
||||
|
||||
Install Quagga on your router, which on most Linuxes is the quagga package. On Debian there is a separate documentation package, quagga-doc. Uncomment this line in `/etc/sysctl.conf` to enable packet forwarding:
|
||||
|
||||
```
|
||||
net.ipv4.ip_forward=1
|
||||
```
|
||||
|
||||
Then run the `sysctl -p` command to load the change.
|
||||
|
||||
### Configuring Quagga
|
||||
|
||||
Look in your Quagga package for example configuration files, such as `/usr/share/doc/quagga/examples/ospfd.conf.sample`. Configuration files should be in `/etc/quagga`, unless your particular Linux flavor does something creative with them. Most Linuxes ship with just two files in this directory, `vtysh.conf` and `zebra.conf`. These provide minimal defaults to enable the daemons to run. zebra always has to run first, and again, unless your distro has done something strange, it should start automatically when you start ospfd. Debian/Ubuntu is a special case, which we will get to in a moment.
|
||||
|
||||
Each router daemon gets its own configuration file, so we must create `/etc/quagga/ospfd.conf`, and populate it with these lines:
|
||||
|
||||
```
|
||||
!/etc/quagga/ospfd.conf
|
||||
hostname router1
|
||||
log file /var/log/quagga/ospfd.log
|
||||
router ospf
|
||||
ospf router-id 192.168.110.15
|
||||
network 192.168.110.0/0 area 0.0.0.0
|
||||
network 192.168.120.0/0 area 0.0.0.0
|
||||
access-list localhost permit 127.0.0.1/32
|
||||
access-list localhost deny any
|
||||
line vty
|
||||
access-class localhost
|
||||
```
|
||||
|
||||
You may use either the exclamation point or hash marks to comment out lines. Let's take a quick walk through these options.
|
||||
|
||||
* **hostname** is whatever you want. This isn't a normal Linux hostname, but the name you see when you log in with `vtysh` or `telnet`.
|
||||
|
||||
* **log file** is whatever file you want to use for the logs.
|
||||
|
||||
* **router** specifies the routing protocol.
|
||||
|
||||
* **ospf router-id** is any 32-bit number. An IP address of the router is good enough.
|
||||
|
||||
* **network** defines the networks your router advertises.
|
||||
|
||||
* The **access-list** entries restrict `vtysh`, the Quagga command shell, to the local machine, and deny remote administration.
|
||||
|
||||
### Debian/Ubuntu
|
||||
|
||||
Debian, Ubuntu, and possibly other Debian derivatives require one more step before you can launch the daemon. Edit `/etc/quagga/daemons` so that all lines say `no` except `zebra`=yes and `ospfd=yes`.
|
||||
|
||||
Then, to launch `ospfd` on Debian launch Quagga:
|
||||
|
||||
```
|
||||
# systemctl start quagga
|
||||
```
|
||||
|
||||
On most other Linuxes, including Fedora and openSUSE, start `ospfd`:
|
||||
|
||||
```
|
||||
# systemctl start ospfd
|
||||
```
|
||||
|
||||
Now Host 1 and Host 2 should ping each other, and the router.
|
||||
|
||||
That was a lot of words to describe a fairly simple setup. In real life the router will connect to two switches and provide a gateway for all the computers attached to those switches. You could add more network interfaces to your router to provide routing for more networks, or connect directly to another router, or to a LAN backbone that connects to other routers.
|
||||
|
||||
You probably don't want to hassle with configuring network interfaces manually. The easy way is to advertise your router with your DHCP server. If you use Dnsmasq then you get DHCP and DNS all in one.
|
||||
|
||||
There are many more configuration options, such as encrypted password protection. See the official documentation at [Quagga Routing Suite][7].
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://www.linux.com/learn/intro-to-linux/2018/3/dynamic-linux-routing-quagga
|
||||
|
||||
作者：[CARLA SCHRODER ][a]
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://www.linux.com/users/cschroder
|
||||
[1]:https://www.linux.com/licenses/category/creative-commons-attribution
|
||||
[2]:https://commons.wikimedia.org/wiki/File:Network_Visualization.png
|
||||
[3]:https://www.linux.com/files/images/networkvisualizationpng
|
||||
[4]:https://www.linux.com/learn/intro-to-linux/2018/2/linux-lan-routing-beginners-part-1
|
||||
[5]:https://www.linux.com/learn/intro-to-linux/2018/3/linux-lan-routing-beginners-part-2
|
||||
[6]:https://www.quagga.net/
|
||||
[7]:https://www.quagga.net/
|
||||
@ -1,108 +0,0 @@
|
||||
Writing Systemd Services for Fun and Profit
|
||||
======
|
||||
|
||||
|
||||
|
||||
Let's say you want to run a games server, a server that runs [Minetest][1], a very cool and open source mining and crafting sandbox game. You want to set it up for your school or friends and have it running on a server in your living room. Because, you know, if that’s good enough for the kernel mailing list admins, then it's good enough for you.
|
||||
|
||||
However, you soon realize it is a chore to remember to run the server every time you switch your computer on and a nuisance to power down safely when you want to switch off.
|
||||
|
||||
First, you have to run the server as a daemon:
|
||||
```
|
||||
minetest --server &
|
||||
|
||||
```
|
||||
|
||||
Take note of the PID (you'll need it later).
|
||||
|
||||
Then you have to tell your friends the server is up by emailing or messaging them. After that you can start playing.
|
||||
|
||||
Suddenly it is 3 am. Time to call it a day! But you can't just switch off your machine and go to bed. First, you have to tell the other players the server is coming down, locate the bit of paper where you wrote the PID we were talking about earlier, and kill the Minetest server gracefully...
|
||||
```
|
||||
kill -2 <PID>
|
||||
|
||||
```
|
||||
|
||||
...because just pulling the plug is a great way to end up with corrupted files. Then and only then can you power down your computer.
|
||||
|
||||
There must be a way to make this easier.
|
||||
|
||||
### Systemd Services to the Rescue
|
||||
|
||||
Let's start off by making a systemd service you can run (manually) as a regular user and build up from there.
|
||||
|
||||
Services you can run without admin privileges live in _~/.config/systemd/user/_ , so start by creating that directory:
|
||||
```
|
||||
cd
|
||||
mkdir -p ~/.config/systemd/user/
|
||||
|
||||
```
|
||||
|
||||
There are several types of systemd _units_ (the formal name of systemd scripts), such as _timers_ , _paths_ , and so on; but what you want is a service. Create a file in _~/.config/systemd/user/_ called _minetest.service_ and open it with your text editor and type the following into it:
|
||||
```
|
||||
# minetest.service
|
||||
|
||||
[Unit]
|
||||
Description= Minetest server
|
||||
Documentation= https://wiki.minetest.net/Main_Page
|
||||
|
||||
[Service]
|
||||
Type= simple
|
||||
ExecStart= /usr/games/minetest --server
|
||||
|
||||
```
|
||||
|
||||
Notice how units have different sections: The `[Unit]` section is mainly informative. It contains information for users describing what the unit is and where you can read more about it.
|
||||
|
||||
The meat of your script is in the `[Service]` section. Here you start by stating what kind of service it is using the `Type` directive. [There are several types][2] of service. If, for example, the process you run first sets up an environment and then calls in another process (which is the main process) and then exits, you would use the `forking` type; if you needed to block the execution of other units until the process in your unit finished, you would use `oneshot`; and so on.
|
||||
|
||||
None of the above is the case for the Minetest server, however. You want to start the server, make it go to the background, and move on. This is what the `simple` type does.
|
||||
|
||||
Next up is the `ExecStart` directive. This directive tells systemd what program to run. In this case, you are going to run `minetest` as headless server. You can add options to your executables as shown above, but you can't chain a bunch of Bash commands together. A line like:
|
||||
```
|
||||
ExecStart: lsmod | grep nvidia > videodrive.txt
|
||||
|
||||
```
|
||||
|
||||
Would not work. If you need to chain Bash commands, it is best to wrap them in a script and execute that.
|
||||
|
||||
Also notice that systemd requires you give the full path to the program. So, even if you have to run something as simple as _ls_ you will have to use `ExecStart= /bin/ls`.
|
||||
|
||||
There is also an `ExecStop` directive that you can use to customize how your service should be terminated. We'll be talking about this directive more in part two, but for now you must know that, if you don't specify an `ExecStop`, systemd will take it on itself to finish the process as gracefully as possible.
|
||||
|
||||
There is a full list of directives in the _systemd.directives_ man page or, if you prefer, [you can check them out on the web][3] and click through to see what each does.
|
||||
|
||||
Although only 6 lines long, your _minetest.service_ is already a fully functional systemd unit. You can run it by executing
|
||||
```
|
||||
systemd --user start minetest
|
||||
|
||||
```
|
||||
|
||||
And stop it with
|
||||
```
|
||||
systemd --user stop minetest
|
||||
|
||||
```
|
||||
|
||||
The `--user` option tells systemd to look for the service in your own directories and to execute the service with your user's privileges.
|
||||
|
||||
That wraps up this part of our server management story. In part two, we’ll go beyond starting and stopping and look at how to send emails to players, alerting them of the server’s availability. Stay tuned.
|
||||
|
||||
Learn more about Linux through the free ["Introduction to Linux" ][4]course from The Linux Foundation and edX.
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://www.linux.com/blog/learn/intro-to-linux/2018/5/writing-systemd-services-fun-and-profit
|
||||
|
||||
作者：[Paul Brown][a]
|
||||
选题：[lujun9972](https://github.com/lujun9972)
|
||||
译者：[译者ID](https://github.com/译者ID)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://www.linux.com/users/bro66
|
||||
[1]:https://www.minetest.net/
|
||||
[2]:http://man7.org/linux/man-pages/man5/systemd.service.5.html
|
||||
[3]:http://man7.org/linux/man-pages/man7/systemd.directives.7.html
|
||||
[4]:https://training.linuxfoundation.org/linux-courses/system-administration-training/introduction-to-linux
|
||||
@ -1,3 +1,5 @@
|
||||
pinewall translating
|
||||
|
||||
Systemd Services: Beyond Starting and Stopping
|
||||
======
|
||||
|
||||
|
||||
@ -1,3 +1,5 @@
|
||||
pinewall translating
|
||||
|
||||
3 Methods To Install Latest Python3 Package On CentOS 6 System
|
||||
======
|
||||
CentOS is RHEL clone and comes with free of cost. It’s a industry standard and cutting edge operating system, this has been used by 90% of webhosting provider since it’s supporting the leading edge server control panel called cPanel/WHM.
|
||||
|
||||
@ -0,0 +1,52 @@
|
||||
机器人学影响 CIO 角色的 3 种方式
|
||||
======
|
||||
|
||||
随着 2017 即将结束，许多 CIO 们的 2018 年目标也将确定。或许你们将参与到机器人流程自动化（RPA）中。多年以来，RPA 对许多公司来说只是一个可望不可及的概念。但是随着组织被迫变得越来越敏捷高效，RPA 所具有的潜在优势开始受到重视。
|
||||
|
||||
根据 Redwood Sofeware 和 Sapio Research 的最新 [研究报告][1]，IT 决策者们相信，未来 5 年有 59% 的业务可以被自动化处理，从而产生新的速度和效率，并且消减相应的重复性的人工工作量的岗位。但是，目前在相应岗位上没有实施 RPA 的公司中，有 20% 的公司员工超过 1000 人。
|
||||
|
||||
对于 CIO 们来说，RPA 会潜在影响你在业务中的角色和你的团队。随着 RPA 的地位在 2018 年中越来越重要，CIO 们和其它 IT 决策者的角色也能会受到如下三种方式的改变：
|
||||
|
||||
### 成为战略性变革代理人的机会增加
|
||||
|
||||
随着压力的增长，用最少的时间做更多的事，内部流程越来越重要。在每个企业中，员工们每天做着既重要又枯燥的一些任务。这些任务可能是枯燥且重复性的任务，但是要求他们必须快速完成且不能出错。
|
||||
|
||||
**[有关你的自动化策略的 ROI 建议，查看我们相关的文章，[如何在自动化方面提升 ROI：4 个小提示][2]。]**
|
||||
|
||||
在财务中从后台的一些操作到采购、供应链、账务、客户服务、以及人力资源，在一个组织中几乎所有的岗位都有一些枯燥的任务。这就为 CIO 们提供了一个机会，将 IT 与业务联合起来，成为使用 RPA 进行战略变革的先锋。
|
||||
|
||||
除了过去的屏幕抓取技术，机器人现在已经实现可定制化，即插即用的解决方案可以根据企业的需要进行设计。使用这种以流程为中心的方法，企业不仅可以将以前由人来完成的任务进行自动化，也可以将应用程序和系统特定任务自动化，比如 ERP 和企业其它应用程序。
|
||||
|
||||
为端到端的流程实施更高级别的自动化将是 CIO 们的价值所在。CIO 们将会站在这种机遇的最前沿。
|
||||
|
||||
### 重新关注人和培训
|
||||
|
||||
技术的变动将让员工更加恐慌，尤其是当这些自动化的变动涉及到他们每日职责的相当大的部分。CIO 们应该清楚地告诉他们，RPA 将如何让他们的角色和责任变得更好，以及用数据说明的、最终将影响到他们底线的战略决策。
|
||||
|
||||
当实施 RPA 时，清晰明确地表达出“人对组织的成功来说是很重要的”信息是很关键的，并且，这种成功需要使技术和人的技能实现适当的平衡。
|
||||
|
||||
CIO 们也应该分析工作流并实施更好的流程，这种流程能够超越以前由终端用户完成的特定任务。通过端到端的流程自动化，CIO 们可以让员工表现更加出色。
|
||||
|
||||
因为在整个自动化的过程中，提升和培训员工技术是很重要的，CIO 们必须与企业高管们一起去决定，帮助员工自信地应对变化的培训项目。
|
||||
|
||||
### 需要长远的规划
|
||||
|
||||
为确保机器人流程自动化的成功，甲方必须有一个长期的方法。这将要求一个弹性的解决方案，它将对整个业务模型有好处，包括消费者。比如，当年亚马逊为他的 Prime 消费者推广快速投递服务时，他们不仅在他们的仓库中重新调整了订单交付流程，他们也自动化了他们的在线客户体验，让它们变得更简化、更快速，并且让消费者比以前更容易下订单。
|
||||
|
||||
在即将到来的一年里，CIO 们可以用同样的方法来应用技术，构建整体的解决方案来改变组织的动作方式。裁员只会让最终效益产生很大影响，但是流程自动化将使 CIO 们能够通过优化和授权来考虑更大的问题。这种方法让 CIO 们去为他们自己和 RPA 构建长期的信誉。这反过来将强化 CIO 们作为探路者的角色，并为整个组织的成功作出贡献。
|
||||
|
||||
对于 CIO 们来说，采用一个长期的、策略性的方法让 RPA 成功，需要时间和艰苦的努力。尽管如此，那些致力于创建一种人力和技术平衡战略的 CIO 们，将会在现在和将来实现更大的价值。
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://enterprisersproject.com/article/2017/11/3-ways-robotics-affects-cio-role
|
||||
|
||||
作者：[Dennis Walsh][a]
|
||||
译者：[qhwdw](https://github.com/qhwdw)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://enterprisersproject.com/user/dennis-walsh
|
||||
[1]:https://www.redwood.com/press-releases/it-decision-makers-speak-59-of-business-processes-could-be-automated-by-2022/
|
||||
[2]:https://enterprisersproject.com/article/2017/11/how-improve-roi-automation-4-tips?sc_cid=70160000000h0aXAAQ
|
||||
111
translated/tech/20180308 Dynamic Linux Routing with Quagga.md
Normal file
111
translated/tech/20180308 Dynamic Linux Routing with Quagga.md
Normal file
@ -0,0 +1,111 @@
|
||||
使用 Quagga 实现 Linux 动态路由
|
||||
============================================================
|
||||
|
||||
|
||||
|
||||
学习如何使用 Quagga 套件的路由协议去管理动态路由。[Creative Commons Attribution][1][Wikimedia Commons: Martin Grandjean][2]
|
||||
|
||||
迄今为止，本系列文章中，我们已经在 [Linux 局域网路由新手指南：第 1 部分][4] 中学习了复杂的 IPv4 地址，在 [Linux 局域网路由新手指南：第 2 部分][5] 中学习了如何去手工创建静态路由。
|
||||
|
||||
今天，我们继续使用 [Quagga][6] 去管理动态路由，这是一个安装完后就不用理它的的软件。Quagga 是一个支持 OSPFv2、OSPFv3、RIP v1 和 v2、RIPng、以及 BGP-4 的路由协议套件，并全部由 zebra 守护程序管理。
|
||||
|
||||
OSPF 的意思是最短路径优先。OSPF 是一个内部网关协议（IGP）;它可以用在局域网和跨因特网的局域网互联中。在你的网络中的每个 OSPF 路由器都包含整个网络的拓扑，并计算通过网络的最短路径。OSPF 会通过多播的方式自动对外传播它检测到的网络变化。你可以将你的网络分割为区域，以保持路由表的可管理性；每个区域的路由器只需要知道离开它的区域的下一跳接口地址，而不用记录你的网络的整个路由表。
|
||||
|
||||
RIP，路由信息协议，是一个很老的协议，RIP 路由器向网络中周期性多播它的整个路由表，而不是像 OSPF 那样只多播网络的变化。RIP 通过跳数来测量路由，任何超过 15 跳的路由它均视为不可到达。RIP 设置很简单，但是 OSPF 在速度、效率、以及弹性方面更佳。
|
||||
|
||||
BGP-4 是边界网关协议版本 4。这是用于因特网流量路由的外部网关协议（EGP）。你不会用到 BGP 协议的，除非你是因特网服务提供商。
|
||||
|
||||
### 准备使用 OSPF
|
||||
|
||||
在我们的小型 KVM 测试实验室中，用两台虚拟机表示两个不同的网络，然后将另一台虚拟机配置为路由器。创建两个网络：net1 是 192.168.110.0/24 而 net2 是 192.168.120.0/24。启用 DHCP 是明智的，否则你要分别进入这三个虚拟机，去为它们设置静态地址。Host 1 在 net1 中，Host 2 在 net2 中，而路由器同时与这两个网络连接。设置 Host 1 的网关地址为 192.168.110.126，Host 2 的网关地址为 192.168.120.136。
|
||||
|
||||
* Host 1： 192.168.110.125
|
||||
|
||||
* Host 2：192.168.120.135
|
||||
|
||||
* Router：192.168.110.126 和 192.168.120.136
|
||||
|
||||
在路由器上安装 Quagga。在大多数 Linux 中它是一个 quagga 包。在 Debian 上还有一个单独的文档包 quagga-doc。取消 `/etc/sysctl.conf` 配置文件中如下这一行的注释去启用包转发功能：
|
||||
|
||||
```
|
||||
net.ipv4.ip_forward=1
|
||||
```
|
||||
|
||||
然后，运行 `sysctl -p` 命令让变化生效。
|
||||
|
||||
### 配置 Quagga
|
||||
|
||||
查看你的 Quagga 包中的示例配置文件，比如，`/usr/share/doc/quagga/examples/ospfd.conf.sample`。除非你的 Linux 版本按你的喜好做了创新，否则，一般情况下配置文件应该在 `/etc/quagga` 目录中。大多数 Linux 版本在这个目录下有两个文件，`vtysh.conf` 和 `zebra.conf`。它们提供了守护程序运行所需要的最小配置。除非你的发行版做了一些特殊的配置，否则，zebra 总是首先运行，当你启动 ospfd 的时候，它将自动启动。Debian/Ubuntu 是一个特例，稍后我们将会说到它。
|
||||
|
||||
每个路由器守护程序将读取它自己的配置文件，因此，我们必须创建 `/etc/quagga/ospfd.conf`，并输入如下内容：
|
||||
|
||||
```
|
||||
!/etc/quagga/ospfd.conf
|
||||
hostname router1
|
||||
log file /var/log/quagga/ospfd.log
|
||||
router ospf
|
||||
ospf router-id 192.168.110.15
|
||||
network 192.168.110.0/0 area 0.0.0.0
|
||||
network 192.168.120.0/0 area 0.0.0.0
|
||||
access-list localhost permit 127.0.0.1/32
|
||||
access-list localhost deny any
|
||||
line vty
|
||||
access-class localhost
|
||||
```
|
||||
|
||||
你可以使用感叹号（！）或者井号（#）去注释掉这些行。我们来快速浏览一下这些选项。
|
||||
|
||||
* **hostname** 是你希望的任何内容。这里不是一般意义上的 Linux 主机名，但是，当你使用 `vtysh` 或者 `telnet` 登入时，你将看到它们。
|
||||
|
||||
* **log file** 是你希望用于保存日志的任意文件。
|
||||
|
||||
* **router** 指定路由协议。
|
||||
|
||||
* **ospf router-id** 是任意的 32 位数字。使用路由器的一个 IP 地址就是很好的选择。
|
||||
|
||||
* **network** 定义你的路由器要通告的网络。
|
||||
|
||||
* **access-list** 限制 `vtysh` 登入，它是 Quagga 命令行 shell，它允许本地机器登入，并拒绝任何远程管理。
|
||||
|
||||
### Debian/Ubuntu
|
||||
|
||||
在你启动守护程序之前，Debian/Ubuntu 相对其它的 Debian 衍生版可能多需要一步到多步。编辑 `/etc/quagga/daemons` ，除了 `zebra`=yes 和 `ospfd=yes` 外，使其它所有的行的值为 `no`。
|
||||
|
||||
然后，在 Debian 上运行 `ospfd` 去启动 Quagga：
|
||||
|
||||
```
|
||||
# systemctl start quagga
|
||||
```
|
||||
|
||||
在大多数的其它 Linux 上，包括 Fedora 和 openSUSE，用如下命令启动 `ospfd`：
|
||||
|
||||
```
|
||||
# systemctl start ospfd
|
||||
```
|
||||
|
||||
现在，Host 1 和 Host 2 将可以互相 ping 通对方和路由器。
|
||||
|
||||
这里用了许多篇幅去描述非常简单的设置。在现实中，路由器将连接两个交换机，然后为连接到这个交换机上的所有电脑提供一个网关。你也可以在你的路由器上添加更多的网络接口，这样你的路由器就可以为更多的网络提供路由服务，或者也可以直接连接到其它路由器上，或者连接到连接其它路由器的骨干网络上。
|
||||
|
||||
你或许不愿意如此麻烦地手工配置网络接口。最简单的方法是使用你的 DHCP 服务器去宣告你的路由器。如果你使用了 Dnsmasq，那么你就有了一个 DHCP 和 DNS 的一体化解决方案。
|
||||
|
||||
还有更多的配置选项，比如，加密的密码保护。更多内容请查看 [Quagga 路由套件][7] 的官方文档。
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://www.linux.com/learn/intro-to-linux/2018/3/dynamic-linux-routing-quagga
|
||||
|
||||
作者：[CARLA SCHRODER ][a]
|
||||
译者：[qhwdw](https://github.com/qhwdw)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://www.linux.com/users/cschroder
|
||||
[1]:https://www.linux.com/licenses/category/creative-commons-attribution
|
||||
[2]:https://commons.wikimedia.org/wiki/File:Network_Visualization.png
|
||||
[3]:https://www.linux.com/files/images/networkvisualizationpng
|
||||
[4]:https://www.linux.com/learn/intro-to-linux/2018/2/linux-lan-routing-beginners-part-1
|
||||
[5]:https://www.linux.com/learn/intro-to-linux/2018/3/linux-lan-routing-beginners-part-2
|
||||
[6]:https://www.quagga.net/
|
||||
[7]:https://www.quagga.net/
|
||||
@ -0,0 +1,106 @@
|
||||
编写有趣且有价值的 Systemd 服务
|
||||
======
|
||||
|
||||
|
||||
|
||||
让我们假设你希望搭建一个游戏服务器，运行 [Minetest][1] 这款非常酷、开源的、以采集 & 合成为主题的沙盒游戏。你希望将游戏运行在位于客厅的服务器中，以便搭建完成后可供你的学校或朋友使用。考虑到内核邮件列表管理就是通过这种方式完成的，那么对你来说也是足够的。
|
||||
|
||||
但你很快发现每次开机之后需要启动服务器，每次关机之前需要安全地关闭服务器，十分繁琐和麻烦。
|
||||
|
||||
最初，你可能用守护进程的方式运行服务器：
|
||||
```
|
||||
minetest --server &
|
||||
|
||||
```
|
||||
记住进程 PID 以便后续使用。
|
||||
|
||||
接着，你还需要通过邮件或短信的方式将服务器已经启动的信息告知你的朋友。然后你就可以开始游戏了。
|
||||
|
||||
转眼之间，已经凌晨三点，今天的战斗即将告一段落。但在你关闭主机、睡个好觉之前，还需要做一些操作。首先，你需要通知其它玩家服务器即将关闭，找到记录我们之前提到的 PID 的纸条，然后友好地关闭 Minetest 服务器。
|
||||
```
|
||||
kill -2 <PID>
|
||||
|
||||
```
|
||||
|
||||
因为直接关闭主机电源很可能导致文件损坏。下一步也是最后一步，关闭主机电源。
|
||||
|
||||
一定有方法能让事情变得更简单。
|
||||
|
||||
### 让 Systemd 服务拯救你
|
||||
|
||||
让我们从构建一个普通用户可以（手动）运行的 systemd 服务开始，然后再逐步增加内容。
|
||||
|
||||
不需要管理员权限即可运行的服务位于 _~/.config/systemd/user/_，故首先需要创建这个目录：
|
||||
```
|
||||
cd
|
||||
mkdir -p ~/.config/systemd/user/
|
||||
|
||||
```
|
||||
有很多类型的 systemd _units_ (曾经叫做 systemd 脚本)，包括 _timers_ 和 _paths_ 等，但我们这里关注的是 service 类型。在 _~/.config/systemd/user/_ 目录中创建 _minetest.service_ 文件，使用文本编辑器打开并输入如下内容：
|
||||
```
|
||||
# minetest.service
|
||||
|
||||
[Unit]
|
||||
Description= Minetest server
|
||||
Documentation= https://wiki.minetest.net/Main_Page
|
||||
|
||||
[Service]
|
||||
Type= simple
|
||||
ExecStart= /usr/games/minetest --server
|
||||
|
||||
```
|
||||
|
||||
可以看到 units 中包含不同的段，其中 `[Unit]` 段主要为用户提供信息，给出 unit 的描述及如何获得更多相关文档。
|
||||
|
||||
脚本核心位于 `[Service]` 段，首先使用 `Type` 指令确定服务类型。服务[有多种类型][2]，下面给出两个示例。如果你运行的进程设置环境变量、调用另外一个进程（主进程）、退出运行，那么你应该使用的服务类型为 `forking`。如果你希望在你的 unit 对应进程结束运行前阻断其他 units 运行，那么你应该使用的服务类型为 `oneshot`。
|
||||
|
||||
但 Minetest 服务器的情形与上面两种都不同，你希望启动服务器并使其在后台持续运行；这种情况下应该使用 `simple` 类型。
|
||||
|
||||
下面来看 `ExecStart` 指令，它给出 systemd 需要运行的程序。在本例中，你希望在后台运行 `minetest` 服务器。如上所示，你可以在可执行程序后面添加参数，但不能将一系列 Bash 命令通过管道连接起来。下面给出的例子无法工作：
|
||||
```
|
||||
ExecStart: lsmod | grep nvidia > videodrive.txt
|
||||
|
||||
```
|
||||
|
||||
如果你需要将 Bash 命令通过管道连接起来，可以将其封装到一个脚本中，然后运行该脚本。
|
||||
|
||||
还需要注意一点，systemd 要求你给出程序的完整路径。故如果你想使用 `simeple` 类型运行类似 _ls_ 的命令，你需要使用 `ExecStart= /bin/ls`。
|
||||
|
||||
另外还有 `ExecStop` 指令用于定制服务终止的方式。我们会在第二部分讨论这个指令，但你要了解，如果你没有指定 `ExecStop`，systemd 会帮你尽可能友好地终止进程。
|
||||
|
||||
_systemd.directives_ 的帮助页中包含完整指令列表，另外你可以在[网站][3]上找到同样的列表，点击即可查看每个指令的具体信息。
|
||||
|
||||
虽然只有 6 行，但你的 _minetest.service_ 已经是一个有完整功能的 systemd unit。执行如下命令启动服务：
|
||||
```
|
||||
systemd --user start minetest
|
||||
|
||||
```
|
||||
|
||||
执行如下命令终止服务
|
||||
```
|
||||
systemd --user stop minetest
|
||||
|
||||
```
|
||||
|
||||
选项 `--user` 告知 systemd 在你的本地目录中检索服务并用你的用户权限执行服务。
|
||||
|
||||
我们的服务器管理故事到此完成了第一部分。在第二部分，我们将在启动和终止服务的基础上，学习如何给用户发邮件、告知用户服务器的可用性。敬请期待。
|
||||
|
||||
可以通过 Linux 基金会和 edX 的免费课程 [Linux 入门][4]学习更多 Linux 知识。
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: https://www.linux.com/blog/learn/intro-to-linux/2018/5/writing-systemd-services-fun-and-profit
|
||||
|
||||
作者：[Paul Brown][a]
|
||||
选题：[lujun9972](https://github.com/lujun9972)
|
||||
译者：[pinewall](https://github.com/pinewall)
|
||||
校对：[校对者ID](https://github.com/校对者ID)
|
||||
|
||||
本文由 [LCTT](https://github.com/LCTT/TranslateProject) 原创编译，[Linux中国](https://linux.cn/) 荣誉推出
|
||||
|
||||
[a]:https://www.linux.com/users/bro66
|
||||
[1]:https://www.minetest.net/
|
||||
[2]:http://man7.org/linux/man-pages/man5/systemd.service.5.html
|
||||
[3]:http://man7.org/linux/man-pages/man7/systemd.directives.7.html
|
||||
[4]:https://training.linuxfoundation.org/linux-courses/system-administration-training/introduction-to-linux
|
||||
Loading…
Reference in New Issue
Block a user