1
00:00:00,000 --> 00:00:10,000
In this video we'll try to understand Linux memory usage.

2
00:00:10,000 --> 00:00:12,000
So how does it work?

3
00:00:12,000 --> 00:00:16,000
Well, a significant portion of RAM in general is used for caching.

4
00:00:16,000 --> 00:00:18,000
And that's convenient.

5
00:00:18,000 --> 00:00:21,000
Your Linux server is reading files from disk

6
00:00:21,000 --> 00:00:26,000
and it's keeping them in RAM as long as the RAM is not needed for anything else.

7
00:00:26,000 --> 00:00:30,000
The benefit is that the next time you need access to the same file

8
00:00:30,000 --> 00:00:36,000
it can be fetched from RAM, which is a lot faster than fetching it from disk.

9
00:00:36,000 --> 00:00:39,000
As a very generic guideline,

10
00:00:39,000 --> 00:00:45,000
a healthy system should have at least 20% of RAM available as free or cached memory.

11
00:00:45,000 --> 00:00:50,000
But there are many exceptions that might justify that you do it differently.

12
00:00:51,000 --> 00:00:56,000
To find current memory usage information, you can use the free utility.

13
00:00:56,000 --> 00:01:00,000
And detailed cache usage can be read from PROC MEMINFO.

14
00:01:00,000 --> 00:01:04,000
If ever you want to optimize memory, you can use SWAP.

15
00:01:04,000 --> 00:01:09,000
As an overflow buffer of emulated RAM on disk where inactive anonymous memory can be parked.

16
00:01:09,000 --> 00:01:12,000
And then there is the out-of-memory killer.

17
00:01:12,000 --> 00:01:15,000
That's a nasty process, part of the Linux kernel,

18
00:01:15,000 --> 00:01:20,000
that will become active and terminate processes on a system that is running low on memory.

19
00:01:20,000 --> 00:01:25,000
There's only one thing that you must understand about the out-of-memory killer.

20
00:01:25,000 --> 00:01:30,000
If ever in your logs you see the ohm killer, you need to add more memory. Period.

21
00:01:30,000 --> 00:01:33,000
Let's check this out a bit.

22
00:01:33,000 --> 00:01:35,000
So here we have 3-M.

23
00:01:35,000 --> 00:01:37,000
And what do we see?

24
00:01:37,000 --> 00:01:41,000
Well, we still have 513 MB of free memory.

25
00:01:41,000 --> 00:01:45,000
And here we can also see that we have 2 GB of available memory.

26
00:01:45,000 --> 00:01:52,000
And available memory is where all the inactive cache is considered as available as well.

27
00:01:52,000 --> 00:01:58,000
And if you want more information about that, well, have a look at slash PROC slash MEMINFO.

28
00:01:58,000 --> 00:02:01,000
Because there you see the active parameters.

29
00:02:01,000 --> 00:02:06,000
And the parameter that really matters here is the inactive file.

30
00:02:06,000 --> 00:02:08,000
That's your inactive cache.

31
00:02:08,000 --> 00:02:10,000
And at the moment more memory is needed.

32
00:02:10,000 --> 00:02:13,000
The inactive cache can be dropped immediately.

33
00:02:13,000 --> 00:02:17,000
Related to that, there's also inactive anonymous memory.

34
00:02:17,000 --> 00:02:19,000
That's your inactive application memory.

35
00:02:19,000 --> 00:02:23,000
And that memory is candidate to being moved to swap.

36
00:02:23,000 --> 00:02:25,000
You learned about that earlier.