mirror of
https://github.com/LCTT/TranslateProject.git
synced 2025-01-25 23:11:02 +08:00
translated
This commit is contained in:
geekpi
parent
a5c1f9419b
commit
526102d38a
@ -1,110 +0,0 @@
|
||||
translating---geekpi
|
||||
|
||||
/dev/[u]random: entropy explained
|
||||
======
|
||||
### Entropy
|
||||
|
||||
When the topic of /dev/random and /dev/urandom come up, you always hear this word: “Entropy”. Everyone seems to have their own analogy for it. So why not me? I like to think of Entropy as “Random juice”. It is juice, required for random to be more random.
|
||||
|
||||
If you have ever generated an SSL certificate, or a GPG key, you may have seen something like:
|
||||
```
|
||||
We need to generate a lot of random bytes. It is a good idea to perform
|
||||
some other action (type on the keyboard, move the mouse, utilize the
|
||||
disks) during the prime generation; this gives the random number
|
||||
generator a better chance to gain enough entropy.
|
||||
++++++++++..+++++.+++++++++++++++.++++++++++...+++++++++++++++...++++++
|
||||
+++++++++++++++++++++++++++++.+++++..+++++.+++++.+++++++++++++++++++++++++>.
|
||||
++++++++++>+++++...........................................................+++++
|
||||
Not enough random bytes available. Please do some other work to give
|
||||
the OS a chance to collect more entropy! (Need 290 more bytes)
|
||||
|
||||
```
|
||||
|
||||
|
||||
By typing on the keyboard, and moving the mouse, you help generate Entropy, or Random Juice.
|
||||
|
||||
You might be asking yourself… Why do I need Entropy? and why it is so important for random to be actually random? Well, lets say our Entropy was limited to keyboard, mouse, and disk IO. But our system is a server, so I know there is no mouse and keyboard input. This means the only factor is your IO. If it is a single disk, that was barely used, you will have low Entropy. This means your systems ability to be random is weak. In other words, I could play the probability game, and significantly decrease the amount of time it would take to crack things like your ssh keys, or decrypt what you thought was an encrypted session.
|
||||
|
||||
Okay, but that is pretty unrealistic right? No, actually it isn’t. Take a look at this [Debian OpenSSH Vulnerability][1]. This particular issue was caused by someone removing some of the code responsible for adding Entropy. Rumor has it they removed it because it was causing valgrind to throw warnings. However, in doing that, random is now MUCH less random. In fact, so much less that Brute forcing the private ssh keys generated is now a fesible attack vector.
|
||||
|
||||
Hopefully by now we understand how important Entropy is to security. Whether you realize you are using it or not.
|
||||
|
||||
### /dev/random & /dev/urandom
|
||||
|
||||
|
||||
/dev/urandom is a Psuedo Random Number Generator, and it **does not** block if you run out of Entropy.
|
||||
/dev/random is a True Random Number Generator, and it **does** block if you run out of Entropy.
|
||||
|
||||
Most often, if we are dealing with something pragmatic, and it doesn’t contain the keys to your nukes, /dev/urandom is the right choice. Otherwise if you go with /dev/random, then when the system runs out of Entropy your application is just going to behave funny. Whether it outright fails, or just hangs until it has enough depends on how you wrote your application.
|
||||
|
||||
### Checking the Entropy
|
||||
|
||||
So, how much Entropy do you have?
|
||||
```
|
||||
[root@testbox test]# cat /proc/sys/kernel/random/poolsize
|
||||
4096
|
||||
[root@testbox test]# cat /proc/sys/kernel/random/entropy_avail
|
||||
2975
|
||||
[root@testbox test]#
|
||||
|
||||
```
|
||||
|
||||
/proc/sys/kernel/random/poolsize, to state the obvious is the size(in bits) of the Entropy Pool. eg: How much random-juice we should save before we stop pumping more. /proc/sys/kernel/random/entropy_avail, is the amount(in bits) of random-juice in the pool currently.
|
||||
|
||||
### How can we influence this number?
|
||||
|
||||
The number is drained as we use it. The most crude example I can come up with is catting /dev/random into /dev/null:
|
||||
```
|
||||
[root@testbox test]# cat /dev/random > /dev/null &
|
||||
[1] 19058
|
||||
[root@testbox test]# cat /proc/sys/kernel/random/entropy_avail
|
||||
0
|
||||
[root@testbox test]# cat /proc/sys/kernel/random/entropy_avail
|
||||
1
|
||||
[root@testbox test]#
|
||||
|
||||
```
|
||||
|
||||
The easiest way to influence this is to run [Haveged][2]. Haveged is a daemon that uses the processor “flutter” to add Entropy to the systems Entropy Pool. Installation and basic setup is pretty straight forward
|
||||
```
|
||||
[root@b08s02ur ~]# systemctl enable haveged
|
||||
Created symlink from /etc/systemd/system/multi-user.target.wants/haveged.service to /usr/lib/systemd/system/haveged.service.
|
||||
[root@b08s02ur ~]# systemctl start haveged
|
||||
[root@b08s02ur ~]#
|
||||
|
||||
```
|
||||
|
||||
On a machine with relatively moderate traffic:
|
||||
```
|
||||
[root@testbox ~]# pv /dev/random > /dev/null
|
||||
40 B 0:00:15 [ 0 B/s] [ <=> ]
|
||||
52 B 0:00:23 [ 0 B/s] [ <=> ]
|
||||
58 B 0:00:25 [5.92 B/s] [ <=> ]
|
||||
64 B 0:00:30 [6.03 B/s] [ <=> ]
|
||||
^C
|
||||
[root@testbox ~]# systemctl start haveged
|
||||
[root@testbox ~]# pv /dev/random > /dev/null
|
||||
7.12MiB 0:00:05 [1.43MiB/s] [ <=> ]
|
||||
15.7MiB 0:00:11 [1.44MiB/s] [ <=> ]
|
||||
27.2MiB 0:00:19 [1.46MiB/s] [ <=> ]
|
||||
43MiB 0:00:30 [1.47MiB/s] [ <=> ]
|
||||
^C
|
||||
[root@testbox ~]#
|
||||
|
||||
```
|
||||
|
||||
Using pv we are able to see how much data we are passing via pipe. As you can see, before haveged, we were getting 2.1 bits per second(B/s). Whereas after starting haveged, and adding processor flutter to our Entropy pool we get ~1.5MiB/sec.
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: http://jhurani.com/linux/2017/11/01/entropy-explained.html
|
||||
|
||||
作者：[James J][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://jblevins.org/log/ssh-vulnkey
|
||||
[1]:http://jhurani.com/linux/2017/11/01/%22https://jblevins.org/log/ssh-vulnkey%22
|
||||
[2]:http://www.issihosts.com/haveged/
|
||||
108
translated/tech/20171101 -dev-urandom- entropy explained.md
Normal file
108
translated/tech/20171101 -dev-urandom- entropy explained.md
Normal file
@ -0,0 +1,108 @@
|
||||
/dev/[u]random：熵解释
|
||||
======
|
||||
### 熵
|
||||
|
||||
当 /dev/random 和 /dev/urandom 的主题出现时，你总是会听到这个词：“熵”。每个人对此似乎都有自己的比喻。那为什么不是我？我喜欢将熵视为“随机果汁”。它是果汁，random 需要它变得更随机。
|
||||
|
||||
如果你曾经生成过 SSL 证书或 GPG 密钥，那么可能已经看到过像下面这样的内容：
|
||||
```
|
||||
We need to generate a lot of random bytes. It is a good idea to perform
|
||||
some other action (type on the keyboard, move the mouse, utilize the
|
||||
disks) during the prime generation; this gives the random number
|
||||
generator a better chance to gain enough entropy.
|
||||
++++++++++..+++++.+++++++++++++++.++++++++++...+++++++++++++++...++++++
|
||||
+++++++++++++++++++++++++++++.+++++..+++++.+++++.+++++++++++++++++++++++++>.
|
||||
++++++++++>+++++...........................................................+++++
|
||||
Not enough random bytes available. Please do some other work to give
|
||||
the OS a chance to collect more entropy! (Need 290 more bytes)
|
||||
|
||||
```
|
||||
|
||||
|
||||
通过在键盘上打字并移动鼠标，你可以帮助生成熵或随机果汁。
|
||||
|
||||
你可能会问自己......为什么我需要熵？以及为什么它对于 random 实际变得随机如此重要？那么，假设我们的熵仅限于键盘、鼠标和磁盘 IO。但是我们的系统是一个服务器，所以我知道没有鼠标和键盘输入。这意味着唯一的因素是你的 IO。如果它是一个单独的几乎不使用磁盘，你将拥有较低的熵。这意味着你的系统随机的能力很弱。换句话说，我可以玩概率游戏，并大幅减少破解 ssh 密钥的时间，或者解密你认为是加密会话的时间。
|
||||
|
||||
好的，但这是非常不切实际的对吧？不，实际上并非如此。看看这个[ Debian OpenSSH 漏洞][1]。这个特定的问题是由于某人删除了一些负责添加熵的代码引起的。有传言说，他们因为它导致 valgrind 发出警告而删除了它。然而，在这样做的时候，random 现在少了很多随机。事实上，熵少了很多因此暴力破解变成了一个可行的攻击向量。
|
||||
|
||||
希望到现在为止，我们理解了熵对安全性的重要性。无论你是否意识到你正在使用它。
|
||||
|
||||
### /dev/random 和 /dev/urandom
|
||||
|
||||
|
||||
/dev/urandom 是一个伪随机随机数生成器，缺乏熵它**不会**停止。
|
||||
/dev/random 是一个真随机数生成器，它会在缺乏熵的时候停止。
|
||||
|
||||
大多数情况下，如果我们正在处理实际的事情，并且它不包含你的核心信息，那么 /dev/urandom 是正确的选择。否则，如果就使用 /dev/random，那么当系统用完熵时，你的程序就会变得有趣。无论它直接失败，或只是挂起，直到它获得足够的熵，这取决于你编写的程序。
|
||||
|
||||
### 检查熵
|
||||
|
||||
那么，你有多少熵？
|
||||
```
|
||||
[root@testbox test]# cat /proc/sys/kernel/random/poolsize
|
||||
4096
|
||||
[root@testbox test]# cat /proc/sys/kernel/random/entropy_avail
|
||||
2975
|
||||
[root@testbox test]#
|
||||
|
||||
```
|
||||
|
||||
/proc/sys/kernel/random/poolsize，说明熵池的大小（以位为单位）。例如：在停止抽水之前我们应该储存多少随机果汁。/proc/sys/kernel/random/entropy_avail 是当前池中随机果汁的数量（以位为单位）。
|
||||
|
||||
### 我们如何影响这个数字？
|
||||
|
||||
这个数字在我们使用时已经耗尽。我可以想出的最简单的例子是将 /dev/random 定向到 /dev/null 中：
|
||||
```
|
||||
[root@testbox test]# cat /dev/random > /dev/null &
|
||||
[1] 19058
|
||||
[root@testbox test]# cat /proc/sys/kernel/random/entropy_avail
|
||||
0
|
||||
[root@testbox test]# cat /proc/sys/kernel/random/entropy_avail
|
||||
1
|
||||
[root@testbox test]#
|
||||
|
||||
```
|
||||
|
||||
影响这个最简单的方法是运行 [Haveged][2]。Haveged 是一个守护进程，它使用处理器的“抖动”将熵添加到系统熵池中。安装和基本设置非常简单。
|
||||
```
|
||||
[root@b08s02ur ~]# systemctl enable haveged
|
||||
Created symlink from /etc/systemd/system/multi-user.target.wants/haveged.service to /usr/lib/systemd/system/haveged.service.
|
||||
[root@b08s02ur ~]# systemctl start haveged
|
||||
[root@b08s02ur ~]#
|
||||
|
||||
```
|
||||
|
||||
在流量相对中等的机器上：
|
||||
```
|
||||
[root@testbox ~]# pv /dev/random > /dev/null
|
||||
40 B 0:00:15 [ 0 B/s] [ <=> ]
|
||||
52 B 0:00:23 [ 0 B/s] [ <=> ]
|
||||
58 B 0:00:25 [5.92 B/s] [ <=> ]
|
||||
64 B 0:00:30 [6.03 B/s] [ <=> ]
|
||||
^C
|
||||
[root@testbox ~]# systemctl start haveged
|
||||
[root@testbox ~]# pv /dev/random > /dev/null
|
||||
7.12MiB 0:00:05 [1.43MiB/s] [ <=> ]
|
||||
15.7MiB 0:00:11 [1.44MiB/s] [ <=> ]
|
||||
27.2MiB 0:00:19 [1.46MiB/s] [ <=> ]
|
||||
43MiB 0:00:30 [1.47MiB/s] [ <=> ]
|
||||
^C
|
||||
[root@testbox ~]#
|
||||
|
||||
```
|
||||
|
||||
使用 pv 我们可以看到我们通过管道传递了多少数据。正如你所看到的，在运行 haveged 之前，我们是 2.1 位/秒（B/s）。而在开始运行 haveged 之后，加入处理器的抖动到我们的熵池中，我们得到大约 1.5MiB/秒。
|
||||
|
||||
--------------------------------------------------------------------------------
|
||||
|
||||
via: http://jhurani.com/linux/2017/11/01/entropy-explained.html
|
||||
|
||||
作者：[James J][a]
|
||||
译者：[geekpi](https://github.com/geekpi)
|
||||
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://jblevins.org/log/ssh-vulnkey
|
||||
[1]:http://jhurani.com/linux/2017/11/01/%22https://jblevins.org/log/ssh-vulnkey%22
|
||||
[2]:http://www.issihosts.com/haveged/
|
||||
Loading…
Reference in New Issue
Block a user