1
00:00:06,750 --> 00:00:07,583
- In this video,

2
00:00:07,583 --> 00:00:10,440
I will show you how to
manage processes using top.

3
00:00:10,440 --> 00:00:11,370
And in order to do so,

4
00:00:11,370 --> 00:00:14,910
I'm starting top with elevated privileges

5
00:00:14,910 --> 00:00:17,610
to ensure that I have
access to everything.

6
00:00:17,610 --> 00:00:18,720
So what do we see?

7
00:00:18,720 --> 00:00:20,160
Well, the top interface

8
00:00:20,160 --> 00:00:21,960
and the top interface is called top

9
00:00:21,960 --> 00:00:25,380
because it's sorts
processes by CPU activity.

10
00:00:25,380 --> 00:00:28,320
So you will always see the busiest process

11
00:00:28,320 --> 00:00:30,540
listed as number one.

12
00:00:30,540 --> 00:00:32,070
In the five headlines,

13
00:00:32,070 --> 00:00:35,340
you get some additional
information about the system.

14
00:00:35,340 --> 00:00:37,620
It starts with some uptime information,

15
00:00:37,620 --> 00:00:40,290
number of users that are logged in.

16
00:00:40,290 --> 00:00:43,170
And then it's giving you a load average.

17
00:00:43,170 --> 00:00:45,030
The load average is giving information

18
00:00:45,030 --> 00:00:48,420
about the current workload of the system.

19
00:00:48,420 --> 00:00:50,940
And it's expressed in three numbers

20
00:00:50,940 --> 00:00:53,493
for the last one, five and 15 minutes.

21
00:00:54,600 --> 00:00:57,960
Now, the number reflects the
average number of processes

22
00:00:57,960 --> 00:01:00,930
that has been asking for attention.

23
00:01:00,930 --> 00:01:02,010
And you know what?

24
00:01:02,010 --> 00:01:04,218
Let's just start a couple of processes.

25
00:01:04,218 --> 00:01:06,801
(quiet typing)

26
00:01:10,860 --> 00:01:13,740
I am going to run three
background processes

27
00:01:13,740 --> 00:01:15,390
because it's more interesting to see

28
00:01:15,390 --> 00:01:18,180
that top is actually showing something.

29
00:01:18,180 --> 00:01:19,020
So as you can see,

30
00:01:19,020 --> 00:01:21,617
we have the three dd processes

31
00:01:21,617 --> 00:01:24,117
that are listed as the
most active processes.

32
00:01:24,117 --> 00:01:27,510
And you can see the load
average slowly climbing up.

33
00:01:27,510 --> 00:01:30,090
It's an average for the last one minute.

34
00:01:30,090 --> 00:01:31,800
So it will take about one minute

35
00:01:31,800 --> 00:01:35,253
until we will see a three right here.

36
00:01:36,270 --> 00:01:38,310
Now, the load average is an indicator

37
00:01:38,310 --> 00:01:40,740
of the performance health of your system.

38
00:01:40,740 --> 00:01:41,970
Big question is,

39
00:01:41,970 --> 00:01:43,860
what is good and what is bad?

40
00:01:43,860 --> 00:01:47,190
Well, it all depends on the
number of CPUs in your system.

41
00:01:47,190 --> 00:01:48,270
And from this interface,

42
00:01:48,270 --> 00:01:51,390
we don't see directly
how many CPUs there are.

43
00:01:51,390 --> 00:01:53,820
The third line is just a
summary of all the CPUs

44
00:01:53,820 --> 00:01:56,373
that are available, but if you press 01,

45
00:01:56,373 --> 00:01:59,610
then you get one line
for every single CPU.

46
00:01:59,610 --> 00:02:01,533
So this is a two CPU system.

47
00:02:02,850 --> 00:02:03,840
So what does that mean?

48
00:02:03,840 --> 00:02:05,232
Well, two CPUs means

49
00:02:05,232 --> 00:02:08,130
that if we have processes

50
00:02:08,130 --> 00:02:11,549
that are asking 100% of CPU attention,

51
00:02:11,549 --> 00:02:16,470
the two CPUs can only have two
processes at the same time.

52
00:02:16,470 --> 00:02:20,460
So if the load average is
higher than the number of CPUs

53
00:02:20,460 --> 00:02:22,530
that you have in your system,

54
00:02:22,530 --> 00:02:25,080
then you need to investigate
what is going on.

55
00:02:25,080 --> 00:02:27,660
It's not necessarily a bad situation

56
00:02:27,660 --> 00:02:30,123
but you should investigate
what is going on.

57
00:02:32,220 --> 00:02:33,180
So the second line,

58
00:02:33,180 --> 00:02:35,760
on the second line we can
see the number of tasks

59
00:02:35,760 --> 00:02:37,050
that is running.

60
00:02:37,050 --> 00:02:38,100
The interesting part.

61
00:02:38,100 --> 00:02:39,540
The interesting part here

62
00:02:39,540 --> 00:02:42,750
is that we can see four of
them are actively running.

63
00:02:42,750 --> 00:02:44,281
Running is for processes

64
00:02:44,281 --> 00:02:47,550
that recently have asked CPU attention.

65
00:02:47,550 --> 00:02:49,800
And we have 244 sleeping,

66
00:02:49,800 --> 00:02:53,070
they are not creating
any significant workload.

67
00:02:53,070 --> 00:02:55,830
No stopped ones, no zombies.

68
00:02:55,830 --> 00:02:57,030
A zombie is a process

69
00:02:57,030 --> 00:02:59,580
that has gone into an unmanageable state

70
00:02:59,580 --> 00:03:01,680
and you shouldn't see them at all.

71
00:03:01,680 --> 00:03:02,730
And if you see them,

72
00:03:02,730 --> 00:03:04,140
well, probably the best thing

73
00:03:04,140 --> 00:03:06,270
is either to wait until they go away

74
00:03:06,270 --> 00:03:09,513
or to reboot your system
to force them to go away.

75
00:03:10,530 --> 00:03:12,090
Then we have the CPUs line.

76
00:03:12,090 --> 00:03:13,380
Remember if I press 01,

77
00:03:13,380 --> 00:03:16,140
then I see a summary for all CPUs

78
00:03:16,140 --> 00:03:19,280
and I can press one again to toggle

79
00:03:19,280 --> 00:03:22,380
for a display that is showing one line

80
00:03:22,380 --> 00:03:24,330
for every single CPU.

81
00:03:24,330 --> 00:03:25,740
In the CPUs line,

82
00:03:25,740 --> 00:03:29,940
we can see information about
what exactly the CPU is doing.

83
00:03:29,940 --> 00:03:32,910
So we can see about 16% in us,

84
00:03:32,910 --> 00:03:35,190
us is user space.

85
00:03:35,190 --> 00:03:37,860
User space are non-privileged processes.

86
00:03:37,860 --> 00:03:41,670
It's typically tasks that are
executed by an ordinary user.

87
00:03:41,670 --> 00:03:45,300
We have 80% in sy, sy is system space.

88
00:03:45,300 --> 00:03:47,940
That is where drivers are involved.

89
00:03:47,940 --> 00:03:50,670
And I can tell you given
the workload of dd,

90
00:03:50,670 --> 00:03:54,900
where we use dd from one
interface to another interface

91
00:03:54,900 --> 00:03:58,320
there it's understandable that
the activity in system space

92
00:03:58,320 --> 00:03:59,910
is quite high, but normally

93
00:03:59,910 --> 00:04:03,303
you should see your activity
more in the user space.

94
00:04:04,230 --> 00:04:08,880
A very important indicator is
id, id is for the idle loop.

95
00:04:08,880 --> 00:04:12,510
It will tell you how much
time your CPU is spending

96
00:04:12,510 --> 00:04:14,130
in doing nothing.

97
00:04:14,130 --> 00:04:17,239
And the other important parameter is a wa,

98
00:04:17,239 --> 00:04:19,560
wa means waiting for I/O

99
00:04:19,560 --> 00:04:22,500
and waiting for I/O means
that your system is waiting

100
00:04:22,500 --> 00:04:26,460
for your disc channel to take
care of processing the data.

101
00:04:26,460 --> 00:04:28,590
If you have a high number in wa

102
00:04:28,590 --> 00:04:30,840
then there is definitely
something not alright

103
00:04:30,840 --> 00:04:33,000
in your storage channel.

104
00:04:33,000 --> 00:04:35,820
And that's a candidate for optimization.

105
00:04:35,820 --> 00:04:38,310
The other parameters are
not so very important.

106
00:04:38,310 --> 00:04:42,000
So keep an eye on us, si, id and wa

107
00:04:42,000 --> 00:04:43,800
and you can get a pretty good impression

108
00:04:43,800 --> 00:04:44,943
of what is going on.

109
00:04:45,840 --> 00:04:48,120
Then we have a line for the memory usage.

110
00:04:48,120 --> 00:04:50,730
So this system as four gigabytes in total

111
00:04:50,730 --> 00:04:54,420
and about 1.2 gigabytes free.

112
00:04:54,420 --> 00:04:58,000
Also, we can see about 1.3 gigabytes used

113
00:04:58,000 --> 00:05:01,800
and 1,400 megabytes in buffer and cache.

114
00:05:01,800 --> 00:05:03,510
Now, what is this buffer and cache?

115
00:05:03,510 --> 00:05:04,680
Simple.

116
00:05:04,680 --> 00:05:06,120
It's all about memory

117
00:05:06,120 --> 00:05:08,727
and this system has four
gigabytes of memory.

118
00:05:08,727 --> 00:05:12,240
Now, Linux is using memory for cache,

119
00:05:12,240 --> 00:05:14,670
if it is not needed by anything else.

120
00:05:14,670 --> 00:05:17,160
Cache is used for
recently requested files.

121
00:05:17,160 --> 00:05:19,560
And the thing is, if the
file is stored in cache,

122
00:05:19,560 --> 00:05:21,630
the next time that somebody needs it,

123
00:05:21,630 --> 00:05:23,280
it can be serviced from cache.

124
00:05:23,280 --> 00:05:25,410
And that makes your system fast.

125
00:05:25,410 --> 00:05:27,750
If the memory is not
needed for anything else,

126
00:05:27,750 --> 00:05:30,090
your data will be kept in cache forever.

127
00:05:30,090 --> 00:05:32,850
At some point, your system
will reach the state

128
00:05:32,850 --> 00:05:35,400
where the memory is
needed for something else.

129
00:05:35,400 --> 00:05:38,550
And then the Linux kernel is
going to clear data from cache,

130
00:05:38,550 --> 00:05:40,773
if it has not recently been used.

131
00:05:41,670 --> 00:05:43,830
Last, we can see the Swap parameter.

132
00:05:43,830 --> 00:05:47,400
The Swap parameter indicates
the usage of Swap space.

133
00:05:47,400 --> 00:05:51,870
Swap space is emulated RAM on disc.

134
00:05:51,870 --> 00:05:54,420
So it's kind of overflow and it is useful

135
00:05:54,420 --> 00:05:57,120
in cases where you are
completely running out of RAM

136
00:05:57,120 --> 00:06:00,150
because then memory pages
can be moved to Swap

137
00:06:00,150 --> 00:06:03,183
and your system can still
continue functioning.

138
00:06:04,350 --> 00:06:06,060
Swap is a little bit slower

139
00:06:06,060 --> 00:06:07,530
but the Linux kernel is dealing

140
00:06:07,530 --> 00:06:09,660
with Swap in a very smart way.

141
00:06:09,660 --> 00:06:13,680
It'll only move inactive
memory pages to Swap,

142
00:06:13,680 --> 00:06:15,090
if that is possible.

143
00:06:15,090 --> 00:06:17,910
And if you only have
inactive memory pages in Swap

144
00:06:17,910 --> 00:06:19,980
then that just cleans up your RAM

145
00:06:19,980 --> 00:06:22,710
and you can use your RAM
for more useful stuff.

146
00:06:22,710 --> 00:06:25,410
And that, in general, is
considered to be good.

147
00:06:25,410 --> 00:06:27,570
As you can see, we have
two gigabytes of Swap

148
00:06:27,570 --> 00:06:29,610
and two gigabytes of Swap are free

149
00:06:29,610 --> 00:06:31,170
and nothing is used.

150
00:06:31,170 --> 00:06:33,570
The last parameter is not about Swap.

151
00:06:33,570 --> 00:06:35,520
This is about a total memory.

152
00:06:35,520 --> 00:06:38,040
So the total memory is set to

153
00:06:38,040 --> 00:06:40,803
more than two gigabytes
of available memory.

154
00:06:41,700 --> 00:06:43,860
Now, the thing that is a
little bit confusing here

155
00:06:43,860 --> 00:06:48,210
is the relation between free
memory and available memory.

156
00:06:48,210 --> 00:06:49,680
Now, what is the relation?

157
00:06:49,680 --> 00:06:51,000
Free memory is memory

158
00:06:51,000 --> 00:06:54,060
that's currently not
used for anything at all.

159
00:06:54,060 --> 00:06:56,940
You won't always see so much free memory

160
00:06:56,940 --> 00:07:01,260
because if your system has
been up for a longer period

161
00:07:01,260 --> 00:07:03,843
then you'll have less
free memory available.

162
00:07:04,680 --> 00:07:07,200
That is because memory will always be used

163
00:07:07,200 --> 00:07:08,970
for caching files.

164
00:07:08,970 --> 00:07:11,460
Now in cache, the Linux kernel keeps track

165
00:07:11,460 --> 00:07:14,880
of memory pages that
are allocated to cache

166
00:07:14,880 --> 00:07:17,190
that have not recently been used.

167
00:07:17,190 --> 00:07:19,710
That's what we call, inactive cache.

168
00:07:19,710 --> 00:07:23,190
And available memory is a
summary of inactive cache

169
00:07:23,190 --> 00:07:24,510
and free memory.

170
00:07:24,510 --> 00:07:26,070
The thing about inactive cache

171
00:07:26,070 --> 00:07:28,800
is that if the memory is
needed by something else,

172
00:07:28,800 --> 00:07:32,310
well, you can just clear
the inactive cache.

173
00:07:32,310 --> 00:07:33,143
And that makes it,

174
00:07:33,143 --> 00:07:35,943
we have 2.3 gigabytes of available memory.

175
00:07:36,780 --> 00:07:39,420
Right in the lowest part
of the top interface,

176
00:07:39,420 --> 00:07:41,370
we can see these processes.

177
00:07:41,370 --> 00:07:43,320
For every process, the PID,

178
00:07:43,320 --> 00:07:45,570
the user that has started to process

179
00:07:45,570 --> 00:07:47,460
and some more information.

180
00:07:47,460 --> 00:07:50,490
The interesting information
right here is the CPU percentage

181
00:07:50,490 --> 00:07:51,600
and the memory.

182
00:07:51,600 --> 00:07:55,440
That's a percentage of CPUs
that the process is running

183
00:07:55,440 --> 00:07:57,990
and the percentage of memory.

184
00:07:57,990 --> 00:08:00,757
And now we can see the
names of these processes.

185
00:08:00,757 --> 00:08:03,090
Let me go out of top for now.

186
00:08:03,090 --> 00:08:04,143
We have seen enough.