salt-syndic是做神马的呢？如果大家知道zabbix proxy的话那就可以很容易理解了，syndic的意思为理事，其实如果叫salt-proxy的话那就更好理解了，它就是一层代理，如同zabbix proxy功能一样，隔离master与minion，使其不需要通讯，只需要与syndic都通讯就可以，这样的话就可以在跨机房的时候将架构清晰部署了，建议zabbix proxy与salt-syndic可以放在一起哦
本次我萌使用node2作为node3的代理让他收到node1（master）的控制
在node1（master）上配置
1 [[email protected] ~]# grep "^[a-Z]" /etc/salt/master 2 default_include: master.d/*.conf 3 file_roots: 4 order_masters: True # 修改这里，表示允许开启多层master
在node2上安装配置
1 [[email protected] salt]# yum install salt-syndic -y 2 [[email protected] salt]# cd /etc/salt/ 3 [[email protected] salt]# grep "^[a-Z]" proxy 4 master: 192.168.56.11 # proxy文件里 5 [[email protected] salt]# grep "^[a-Z]" master 6 syndic_master: 192.168.56.11 # master文件里 7 [[email protected] salt]# systemctl start salt-master.service 8 [[email protected] salt]# systemctl start salt-syndic.service 9 [[email protected] salt]# netstat -tpln 10 Active Internet connections (only servers) 11 Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name 12 tcp 0 0 0.0.0.0:111 0.0.0.0:* LISTEN 1/systemd 13 tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN 998/sshd 14 tcp 0 0 0.0.0.0:4505 0.0.0.0:* LISTEN 6013/python 15 tcp 0 0 0.0.0.0:4506 0.0.0.0:* LISTEN 6019/python 16 tcp6 0 0 :::111 :::* LISTEN 1/systemd 17 tcp6 0 0 :::22 :::* LISTEN 998/sshd
node3上正常安装minion
1 [[email protected] salt]# yum install salt-minion -y 2 [[email protected] salt]# cd /etc/salt/ 3 [[email protected] salt]# grep "^[a-Z]" minion 4 master: 192.168.56.12 # 此时只需要认定node2就好，不需要知道node1的存在 5 [[email protected] salt]# systemctl start salt-minion
然后回到node2（syndic）
1 [[email protected] salt]# salt-key -L 2 Accepted Keys: 3 Denied Keys: 4 Unaccepted Keys: 5 linux-node3.example.com 6 Rejected Keys: 7 [[email protected] salt]# salt-key –A # 把key接受了
最后回到node1（master）
1 [[email protected] ~]# salt-key –L # 发现并没有linux-node3.example.com 2 Accepted Keys: 3 linux-node1.example.com 4 linux-node2.example.com 5 Denied Keys: 6 Unaccepted Keys: 7 Rejected Keys: 8 [[email protected] ~]# salt ‘*‘ test.ping 9 linux-node2.example.com: 10 True 11 linux-node1.example.com: 12 True 13 linux-node3.example.com: # 但是它会出现效果 14 True
其他的同层代理及多层代理的配置也是相同的，只需要分清每个代理的上层master就好
这里有一些常见的问题
1.我代理之下控制的是否可以重名？举个例子就是node3的id改成node2，然后在总master上执行会有什么情况？
首先我萌要涉及到修改id啦，小伙伴还记得修改id的流程吗~~
1 [[email protected] salt]# systemctl stop salt-minion.service # 停止minion 2 [[email protected] salt]# salt-key –L # 注意是在node2（syndic）上哦，因为node3的眼里的master是node2，并且把key是发送给node2了哦，删除它 3 Accepted Keys: 4 linux-node3.example.com 5 Denied Keys: 6 Unaccepted Keys: 7 Rejected Keys: 8 [[email protected] salt]# salt-key -d linux-node3.example.com 9 The following keys are going to be deleted: 10 Accepted Keys: 11 linux-node3.example.com 12 Proceed? [N/y] y 13 Key for minion linux-node3.example.com deleted. 14 [[email protected] salt]# rm -fr /etc/salt/pki/minion/ # 删除minion端/etc/salt/pki/minion下所有文件 15 [[email protected] salt]# grep "^[a-Z]" /etc/salt/minion # 修改新id 16 master: 192.168.56.12 17 id: linux-node2.example.com # 配置一个已有的重复id做测试 18 [[email protected] salt]# systemctl start salt-minion.service # 再次启动minion 19 [[email protected] salt]# salt-key –L # 回到node2再次接受新id的key 20 Accepted Keys: 21 Denied Keys: 22 Unaccepted Keys: 23 linux-node2.example.com 24 Rejected Keys: 25 [[email protected] salt]# salt-key -A 26 The following keys are going to be accepted: 27 Unaccepted Keys: 28 linux-node2.example.com 29 Proceed? [n/Y] Y 30 Key for minion linux-node2.example.com accepted. 31 [[email protected] salt]# salt ‘*‘ test.ping # 简单测试下 32 linux-node2.example.com: 33 True 34 35 最后验证我们的测试，回到node1（master） 36 [[email protected] ~]# salt ‘*‘ test.ping 37 linux-node2.example.com: 38 True 39 linux-node1.example.com: 40 True 41 linux-node2.example.com: 42 True 43 我萌发现，wtf，什么鬼，linux-node2.example.com居然出现了两次，虽然已经想过这种情况，但是在实际使用中我肯定是分不清谁是谁了，所以这种使用了代理后依然id重名的方式依然是很不好的，所以还是建议大家把id要分清楚哦，最简单的方式就是设置合理的主机名，这样所有的机器都不会重复，而且连设置id这个事情都可以省略了（我已经将node3的id改回去了）
2.远程执行没有问题了，这种架构下状态文件的执行会不会有影响呢？
1 [[email protected] base]# pwd # 我们在master上定义top 2 /srv/salt/base 3 [[email protected] base]# cat top.sls # 其实就是给大家传输了个文件 4 base: 5 ‘*‘: 6 - known-hosts.known-hosts 7 [[email protected] base]# cat known-hosts/known-hosts.sls 8 known-hosts: 9 file.managed: 10 - name: /root/.ssh/known_hosts 11 - source: salt://known-hosts/templates/known-hosts 12 - clean: True 13 [[email protected] base]# salt ‘*‘ state.highstate 14 linux-node3.example.com: 15 ---------- 16 ID: states 17 Function: no.None 18 Result: False 19 Comment: No Top file or master_tops data matches found. 20 Changes: 21 22 Summary for linux-node3.example.com 23 ------------ 24 Succeeded: 0 25 Failed: 1 26 ------------ 27 Total states run: 1 28 Total run time: 0.000 ms 29 linux-node2.example.com: 30 ---------- 31 ID: known-hosts 32 Function: file.managed 33 Name: /root/.ssh/known_hosts 34 Result: True 35 Comment: File /root/.ssh/known_hosts updated 36 Started: 11:15:35.210699 37 Duration: 37.978 ms 38 Changes: 39 ---------- 40 diff: 41 New file 42 mode: 43 0644 44 45 Summary for linux-node2.example.com 46 ------------ 47 Succeeded: 1 (changed=1) 48 Failed: 0 49 ------------ 50 Total states run: 1 51 Total run time: 37.978 ms 52 linux-node1.example.com: 53 ---------- 54 ID: known-hosts 55 Function: file.managed 56 Name: /root/.ssh/known_hosts 57 Result: True 58 Comment: File /root/.ssh/known_hosts is in the correct state 59 Started: 11:15:35.226119 60 Duration: 51.202 ms 61 Changes: 62 63 Summary for linux-node1.example.com 64 ------------ 65 Succeeded: 1 66 Failed: 0 67 ------------ 68 Total states run: 1 69 Total run time: 51.202 ms 70 ERROR: Minions returned with non-zero exit code 71 显而易见的node3发生了错误，而node1跟node2正常（很好理解），我们去看node3报出的“No Top file or master_tops data matches found”，言简意赅没有找到匹配的top执行文件，简单推断出是因为node3认证的master是node2，但是node2上没有写top，我们去node2上写一个不同的top再次测试下 72 [[email protected] base]# pwd 73 /srv/salt/base 74 [[email protected] base]# cat top.sls # 这个更简单了，就是ls /root 75 base: 76 ‘*‘: 77 - cmd.cmd 78 [[email protected] base]# cat cmd/cmd.sls 79 cmd: 80 cmd.run: 81 - name: ls /root 82 好的我们回到master上再次测试，我将node1、2正常执行的信息省略 83 [[email protected] base]# salt ‘*‘ state.highstate 84 linux-node3.example.com: 85 ---------- 86 ID: cmd 87 Function: cmd.run 88 Name: ls /root 89 Result: True 90 Comment: Command "ls /root" run 91 Started: 11:24:42.752326 92 Duration: 11.944 ms 93 Changes: 94 ---------- 95 pid: 96 5095 97 retcode: 98 0 99 stderr: 100 stdout: 101 lvs.sh 102 103 Summary for linux-node3.example.com 104 ------------ 105 Succeeded: 1 (changed=1) 106 Failed: 0 107 ------------ 108 Total states run: 1 109 Total run time: 11.944 ms 110 我们已经可以看出一些端倪，我们再次修改master的配置文件并执行测试 111 [[email protected] base]# cat top.sls 112 base: 113 ‘linux-node3.example.com‘: # 只定义执行node3 114 - known-hosts.known-hosts 115 [[email protected] base]# salt ‘*‘ state.highstate 116 linux-node3.example.com: 117 ---------- 118 ID: cmd 119 Function: cmd.run 120 Name: ls /root 121 Result: True 122 Comment: Command "ls /root" run 123 Started: 11:28:20.792475 124 Duration: 8.686 ms 125 Changes: 126 ---------- 127 pid: 128 5283 129 retcode: 130 0 131 stderr: 132 stdout: 133 lvs.sh 134 135 Summary for linux-node3.example.com 136 ------------ 137 Succeeded: 1 (changed=1) 138 Failed: 0 139 ------------ 140 Total states run: 1 141 Total run time: 8.686 ms 142 linux-node2.example.com: 143 ---------- 144 ID: states 145 Function: no.None 146 Result: False 147 Comment: No Top file or master_tops data matches found. 148 Changes: 149 150 Summary for linux-node2.example.com 151 ------------ 152 Succeeded: 0 153 Failed: 1 154 ------------ 155 Total states run: 1 156 Total run time: 0.000 ms 157 linux-node1.example.com: 158 ---------- 159 ID: states 160 Function: no.None 161 Result: False 162 Comment: No Top file or master_tops data matches found. 163 Changes: 164 165 Summary for linux-node1.example.com 166 ------------ 167 Succeeded: 0 168 Failed: 1 169 ------------ 170 Total states run: 1 171 Total run time: 0.000 ms 172 ERROR: Minions returned with non-zero exit code 173 我们发现这次node1跟node2出刚才问题了，而node3执行的是node2上定义的top，好吧，这时候就要发挥小北方的作用！ 174 北方的总结： 175 每个minion会去找自己master里定义的top并执行，即node1、2找的是master的，而node2找的是syndic（node2）的 176 177 “No Top file or master_tops data matches found”出现是因为我每次执行都是salt ‘*‘ state.highstate，即让所有机器都查找top文件并执行对应操作，第一次node3出现问题是因为它听从的top文件在syndic上，当时syndic上我还没有写top所以他找不到匹配自己的；第二次我把top里执行的*换成了node3单独一个，没有node1跟node2的相关操作了，他们接受到指令并来查找top文件想执行相关操作发现没匹配自己也因此报错，就跟刚才node3找不到是一个意思 178 179 一下子还是无法理解呢，那么怎么办呢，有一个规范的做法就是，将master的文件目录直接拷到所有的syndic上，这样就可以保证所有的操作都是统一的了，如同没有代理的时候一样
3.top好麻烦呀，那么我普通的执行sls文件会怎么样呢？
1 [[email protected] base]# salt ‘*‘ state.sls known-hosts.known-hosts 2 linux-node3.example.com: 3 Data failed to compile: 4 ---------- 5 No matching sls found for ‘known-hosts.known-hosts‘ in env ‘base‘ 6 linux-node2.example.com: 7 ---------- 8 ID: known-hosts 9 Function: file.managed 10 Name: /root/.ssh/known_hosts 11 Result: True 12 Comment: File /root/.ssh/known_hosts is in the correct state 13 Started: 11:46:03.968021 14 Duration: 870.596 ms 15 Changes: 16 17 Summary for linux-node2.example.com 18 ------------ 19 Succeeded: 1 20 Failed: 0 21 ------------ 22 Total states run: 1 23 Total run time: 870.596 ms 24 linux-node1.example.com: 25 ---------- 26 ID: known-hosts 27 Function: file.managed 28 Name: /root/.ssh/known_hosts 29 Result: True 30 Comment: File /root/.ssh/known_hosts is in the correct state 31 Started: 11:46:05.003462 32 Duration: 42.02 ms 33 Changes: 34 35 Summary for linux-node1.example.com 36 ------------ 37 Succeeded: 1 38 Failed: 0 39 ------------ 40 Total states run: 1 41 Total run time: 42.020 ms 42 ERROR: Minions returned with non-zero exit code 43 我么看到node3又报错了，“No matching sls found for ‘known-hosts.known-hosts‘ in env ‘base‘”，我甚至都不需要验证都知道是怎么回事了，直接复制下来 44 45 每个minion会去找自己master里定义的sls并执行，即node1、2找的是master的，而node2找的是syndic（node2）的 46 47 所以如果你在syndic定义个known-hosts但是里面执行些其他操作那么node3就会按这个来了，但是没有人会这么乱七八糟的搞，因此：保证各个syndic与master的文件目录保持统一！
时间： 2024-11-06 03:40:27