文章在我的51cto博客上也转了一份:
http://rfyiamcool.blog.51cto.com/1030776/1433843
这两天有人问我saltstack 关于mine的主要应用场景和具体的用法,我翻了下自己一些写的关于saltstack的博客,还真没有mine的文章。。。 现在总结下mine的文档,供大家拍砖。
mine是做什么? 就是你在mine定义了收集模块,他自己会把数据收集的,存到了哪里,不知道,但是肯定是客户端minino本地。有兴趣的朋友可以check saltstack的代码。
这里先粘贴下,官方的doc用法:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
|
mine.delete:
Remove specific function contents of minion. Returns True on success.
CLI Example:
salt ‘*’ mine.delete ‘network.interfaces’
mine.flush:
Remove all mine contents of minion. Returns True on success.
CLI Example:
salt ‘*’ mine.flush
mine.get:
Get data from the mine based on the target, function and expr_form
Targets can be matched based on any standard matching system that can be
matched on the master via these keywords::
glob
pcre
grain
grain_pcre
CLI Example:
salt ‘*’ mine.get ‘*’ network.interfaces
salt ‘*’ mine.get ‘os:Fedora’ network.interfaces grain
mine.send:
Send a specific function to the mine.
CLI Example:
salt ‘*’ mine.send network.interfaces eth0
mine.update: 原文:
Execute the configured functions and send the data back up to the master
The functions to be executed are merged from the master config, pillar and
minion config under the option “function_cache”:
mine_functions:
network.ip_addrs:
|
配置起来很简单,跑到minion端,配置下文件,开启几个模块的检测。
[root@vm-10-154-252-46 utils]# tail /etc/salt/minion
mine_functions:
test.ping: []
network.dig:
– www.naver.com
network.usage: []
mine_interval: 1
然后咱们跑到master端去抓数据。
[root@vm-10-154-252-82 ~/clusterops/templates]$salt ‘10.154.252.46’ mine.get ‘*’ network.dig
10.154.252.46:
———-
10.154.252.46:
; <<>> DiG 9.8.2rc1-RedHat-9.8.2-0.10.rc1.el6 <<>> www.baidu.com
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 5279
;; flags: qr rd ra; QUERY: 1, ANSWER: 3, AUTHORITY: 5, ADDITIONAL: 2
;; QUESTION SECTION:
;www.baidu.com. IN A
;; ANSWER SECTION:
www.baidu.com. 186 IN CNAME www.a.shifen.com.
www.a.shifen.com. 149 IN A 220.181.111.188
www.a.shifen.com. 149 IN A 220.181.112.244
;; AUTHORITY SECTION:
a.shifen.com. 186 IN NS ns3.a.shifen.com.
a.shifen.com. 186 IN NS ns1.a.shifen.com.
a.shifen.com. 186 IN NS ns5.a.shifen.com.
a.shifen.com. 186 IN NS ns2.a.shifen.com.
a.shifen.com. 186 IN NS ns4.a.shifen.com.
;; ADDITIONAL SECTION:
ns5.a.shifen.com. 70 IN A 119.75.219.43
ns2.a.shifen.com. 186 IN A 180.149.133.241
;; Query time: 1 msec
;; SERVER: 10.150.0.253#53(10.150.0.253)
;; WHEN: Wed Jul 2 14:42:55 2014
;; MSG SIZE rcvd: 212
原文: http://rfyiamcool.blog.51cto.com/1030776/1433843
[root@vm-10-154-252-82 ~/clusterops/templates]$
我先前定义的是一分钟来生成一次数据。在minon的日志端是可以体现出来的。
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
|
2014-07-03 09:15:05,054 [salt.loaded.int.module.mine
2014-07-03 09:15:05,186 [salt.loaded.int.module.mine
2014-07-03 09:16:05,149 [salt.loaded.int.module.mine
2014-07-03 09:16:05,274 [salt.loaded.int.module.mine
2014-07-03 09:17:05,242 [salt.loaded.int.module.mine
2014-07-03 09:17:05,366 [salt.loaded.int.module.mine
2014-07-03 09:18:05,336 [salt.loaded.int.module.mine
2014-07-03 09:18:05,463 [salt.loaded.int.module.mine
2014-07-03 09:19:05,430 [salt.loaded.int.module.mine
2014-07-03 09:19:05,554 [salt.loaded.int.module.mine
2014-07-03 09:20:05,523 [salt.loaded.int.module.mine
2014-07-03 09:20:05,648 [salt.loaded.int.module.mine
2014-07-03 09:21:05,619 [salt.loaded.int.module.mine
2014-07-03 09:21:05,748 [salt.loaded.int.module.mine
2014-07-03 09:22:05,713 [salt.loaded.int.module.mine
2014-07-03 09:22:05,843 [salt.loaded.int.module.mine
2014-07-03 09:23:05,808 [salt.loaded.int.module.mine
2014-07-03 09:23:05,938 [salt.loaded.int.module.mine
2014-07-03 09:24:05,901 [salt.loaded.int.module.mine
2014-07-03 09:24:06,026 [salt.loaded.int.module.mine
2014-07-03 09:25:05,995 [salt.loaded.int.module.mine
2014-07-03 09:25:06,120 [salt.loaded.int.module.mine
2014-07-03 09:26:05,088 [salt.loaded.int.module.mine
2014-07-03 09:26:05,214 [salt.loaded.int.module.mine
2014-07-03 09:27:05,183 [salt.loaded.int.module.mine
2014-07-03 09:27:05,319 [salt.loaded.int.module.mine
2014-07-03 09:28:05,287 [salt.loaded.int.module.mine
2014-07-03 09:28:05,418 [salt.loaded.int.module.mine
2014-07-03 09:29:05,373 [salt.loaded.int.module.mine
2014-07-03 09:29:05,502 [salt.loaded.int.module.mine
2014-07-03 09:30:05,466 [salt.loaded.int.module.mine
2014-07-03 09:30:05,591 [salt.loaded.int.module.mine
2014-07-03 09:31:05,560 [salt.loaded.int.module.mine
2014-07-03 09:31:05,689 [salt.loaded.int.module.mine
2014-07-03 09:32:05,655 [salt.loaded.int.module.mine
2014-07-03 09:32:05,785 [salt.loaded.int.module.mine
2014-07-03 09:33:05,749 [salt.loaded.int.module.mine
2014-07-03 09:33:05,874 [salt.loaded.int.module.mine
2014-07-03 09:34:05,845 [salt.loaded.int.module.mine
2014-07-03 09:34:05,969 [salt.loaded.int.module.mine
2014-07-03 09:35:05,937 [salt.loaded.int.module.mine
2014-07-03 09:35:06,068 [salt.loaded.int.module.mine
2014-07-03 09:36:05,017 [salt.loaded.int.module.mine
2014-07-03 09:36:05,141 [salt.loaded.int.module.mine
2014-07-03 09:37:05,113 [salt.loaded.int.module.mine
2014-07-03 09:37:05,238 [salt.loaded.int.module.mine
2014-07-03 09:38:05,208 [salt.loaded.int.module.mine
2014-07-03 09:38:05,336 [salt.loaded.int.module.mine
2014-07-03 09:39:05,304 [salt.loaded.int.module.mine
2014-07-03 09:39:05,428 [salt.loaded.int.module.mine
2014-07-03 09:40:05,687 [salt.loaded.int.module.mine
2014-07-03 09:40:05,810 [salt.loaded.int.module.mine
2014-07-03 09:41:05,781 [salt.loaded.int.module.mine
2014-07-03 09:41:05,908 [salt.loaded.int.module.mine
|
咱们也可以利用mine.send来发送数据。
[root@vm-10-154-252-82 ~/clusterops/templates]$salt ‘10.154.252.46’ mine.send network.dig youku.com
10.154.252.46:
True
[root@vm-10-154-252-82 ~/clusterops/templates]$
我们又重新定义了mine的规则。
1
2
3
4
|
mine_functions:
test.ping: []
cmd.run:
– date;df
|
咱们定义了有时间特征的收集,用来判断下,他获取的数据是不是一分钟间隔,你就算mine.get也只是获取到他先前生成的数据,而不是实时的。
原文:http://rfyiamcool.blog.51cto.com/1030776/1433843
有些朋友还是再疑惑这些东西也可以用直接推送模块来实现,但是你看下面,我定义了sleep 10s,如果用模块去取值的话,一定要sleep 10秒之后,才能获取到数据。但这里master去get数据的时候,还是很即时的回来。当然取得数据是一分钟之前的。对一些场合这已经足够了。
上面说了,是什么场合? 来个简单渲染haproxy配置。
/srv/pillar/top.sls:
base:
‘G@roles:web’:
– web
原文:http://rfyiamcool.blog.51cto.com/1030776/1433843
/srv/pillar/web.sls:
mine_functions:
network.ip_addrs: [eth0]
/etc/salt/minion.d/mine.conf:
mine_interval: 5
/srv/salt/haproxy.sls:
haproxy_config:
file:
– managed
– name: /etc/haproxy/config
– source: salt://haproxy_config
– template: jinja
/srv/salt/haproxy_config:
<…file contents snipped…>
{% for server, addrs in salt[‘mine.get’](‘roles:web’, ‘network.ip_addrs’, expr_form=’grain’).items() %}
server {{ server }} {{ addrs[0] }}:80 check
{% endfor %}
<…file contents snipped…>
根据monin客户端的一些特征来渲染配置,估计大家已经学会了用根据grains取渲染配置,但是grains的数据有些不及时,虽然也有办法让他及时。
我这里也只是阐述下mine和grains在通过jinja2渲染配置的场景。 我自己看来,grains更适合很久很久才变化的。 mine适合变化比较大的。
大家show下saltstack grains的数据,几乎都是系统和硬件的信息。
1
2
3
4
5
6
7
8
9
10
11
12
13
|
[root@vm-10-154-252-82 ~/clusterops]$salt ‘10.154.252.46’ grains.item productname
10.154.252.46:
productname: OpenStack Nova
[root@vm-10-154-252-82 ~/clusterops]$salt ‘10.154.252.46’ grains.item lang
10.154.252.46:
lang: python erlang
[root@vm-10-154-252-82 ~/clusterops]$
[root@vm-10-154-252-82 ~/clusterops]$
[root@vm-10-154-252-82 ~/clusterops]$
[root@vm-10-154-252-82 ~/clusterops]$salt ‘10.154.252.46’ sys.reload_modules
10.154.252.46:
True
[root@vm-10-154-252-82 ~/clusterops]$
|
用mine的话,咱们可以定义自己感兴趣的。 可以根据现在minon使用的负载和内存情况,来分配不同的配置选项。。。。 pillar 和 mine我想大家应该能区分开来 。
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
|
[root@vm-10-154-252-82 ~/clusterops]$time salt ‘10.154.252.46’ mine.get ‘*’ status.meminfo|more
10.154.252.46:
———-
10.154.252.46:
———-
Active:
———-
unit:
kB
value:
506876
Active(anon):
———-
unit:
kB
value:
94456
Active(file):
———-
unit:
kB
value:
412420
AnonHugePages:
———-
unit:
kB
value:
20480
AnonPages:
———-
unit:
kB
value:
94232
Bounce:
———-
|
总结,一定要区分直接推送模块和mine的区别,还有grains和mine的区别。 仁者见仁,智者见智。 想怎么用,还是需要你自己掌控的。