swarm集群应用操作

avatar 2020年3月14日18:04:05 评论 1,391 次浏览

swarm集群服务创建网络

集群的基本操作熟悉了,我们就在集群上部署个应用,然后看看swarm集群中的应用操作方法。在swarm集群中创建应用之前需要先创建一个网络,因为swarm集群中的节点是依靠网络来管理节点之间的通信的。

[root@wulaoer.org ~]# docker network create -d overlay wulaoer_net
[root@wulaoer.org  ~]# docker network ls
NETWORK ID          NAME                DRIVER              SCOPE
a38745e6b5b8        bridge              bridge              local
a01287537f07        docker_gwbridge     bridge              local
09784c5005ab        host                host                local
pddr4assraqs        ingress             overlay             swarm
fe6de3090b26        none                null                local
iwsgekycuzsb        wulaoer_net         overlay             swarm

swarm集群中创建应用

网络创建之后才可以创建应用,下面我们创建一个nginx应用。

[root@wulaoer.org  ~]# docker service create --replicas 1 --network wulaoer_net --name wulaoer_web -p 80:80 nginx
vlm6qznhp54qma1up7xeuzdq2
overall progress: 1 out of 1 tasks 
1/1: running   [==================================================>] 
verify: Service converged 

nginx服务已经创建完成,我们先看一下nginx是运行在那个节点上面

[root@wulaoer.org  ~]# docker service ps wulaoer_web
ID                  NAME                IMAGE               NODE                    DESIRED STATE       CURRENT STATE            ERROR               PORTS
nja5m0crefrv        wulaoer_web.1       nginx:latest        slave   Running             Running 10 minutes ago                       

只运行了一个副本,运行在slave节点上,我们测试一下swarm是否能够做节点转发,我们使用curl访问三个节点的80端口,获取状态值。

[root@wulaoer.org  ~]# curl -I -m 10 -o /dev/null -s -w %{http_code} 10.211.55.143
200[root@wulaoer.org  ~]# curl -I -m 10 -o /dev/null -s -w %{http_code} 10.211.55.144
200[root@wulaoer.org  ~]# curl -I -m 10 -o /dev/null -s -w %{http_code} 10.211.55.145
200[root@wulaoer.org  ~]# 

swarm集群是可以做节点转发的,这里尝试把master节点的docker关闭,其他节点的应用不影响访问。

应用自动转移

我们尝试一下扩展多个容器,扩展不是在原来的基础上扩展,而是扩展后的总数,这里千万要记得。

[root@wulaoer.org  ~]# docker service scale wulaoer_web=5  #这里的5是总副本数,如果想启动10个副本必须把5改成10,不能在启动一下,不会在五的基础上在加5
wulaoer_web scaled to 5
overall progress: 5 out of 5 tasks 
1/5: running   [==================================================>] 
2/5: running   [==================================================>] 
3/5: running   [==================================================>] 
4/5: running   [==================================================>] 
5/5: running   [==================================================>] 
verify: Service converged 

我们看一下增加的五个容器节点运行在哪个服务器上了。

[root@wulaoer.org  ~]# docker service ps wulaoer_web
ID                  NAME                IMAGE               NODE                    DESIRED STATE       CURRENT STATE            ERROR               PORTS
nja5m0crefrv        wulaoer_web.1       nginx:latest        slave  					 Running             Running 27 minutes ago                       
8wazfd9562tj        wulaoer_web.2       nginx:latest        wulaoer.org              Running             Running 3 minutes ago                        
kmntagm4lgk0        wulaoer_web.3       nginx:latest        slave   				 Running             Running 5 minutes ago                        
pd8w2f9ivivk        wulaoer_web.4       nginx:latest        node                     Running             Running 3 minutes ago                        
nsld1eq9qoka        wulaoer_web.5       nginx:latest        node                     Running             Running 3 minutes ago                        

随机分配所有节点上都有增加的容器了,我们尝试一下,下线一个节点,五个容器应用是否有什么变化

[root@wulaoer.org  ~]# docker node update --availability drain mrtks5bwhpe5k3vskv2kctf66
mrtks5bwhpe5k3vskv2kctf66
[root@wulaoer.org  ~]# docker service ps wulaoer_web
ID                  NAME                IMAGE               NODE                    DESIRED STATE       CURRENT STATE            ERROR               PORTS
q98hcjj9id87        wulaoer_web.1       nginx:latest        wulaoer.org              Running             Running 1 second ago                         
nja5m0crefrv         \_ wulaoer_web.1   nginx:latest        slave                    Shutdown            Shutdown 2 seconds ago                       
8wazfd9562tj        wulaoer_web.2       nginx:latest        wulaoer.org              Running             Running 6 minutes ago                        
xqyjcwlznjmy        wulaoer_web.3       nginx:latest        node                     Running             Running 1 second ago                         
kmntagm4lgk0         \_ wulaoer_web.3   nginx:latest        slave                    Shutdown            Shutdown 2 seconds ago                       
pd8w2f9ivivk        wulaoer_web.4       nginx:latest        node                     Running             Running 6 minutes ago                        
nsld1eq9qoka        wulaoer_web.5       nginx:latest        node                     Running             Running 6 minutes ago  

节点下线后自动转到别的节点上了,我们尝试着上线节点,看看是否会重新分配到nginx的应用。

[root@wulaoer.org  ~]# docker node update --availability active mrtks5bwhpe5k3vskv2kctf66
mrtks5bwhpe5k3vskv2kctf66
[root@wulaoer.org  ~]# docker service ps wulaoer_web
ID                  NAME                IMAGE               NODE                    DESIRED STATE       CURRENT STATE            ERROR               PORTS
q98hcjj9id87        wulaoer_web.1       nginx:latest        wulaoer.org              Running             Running 2 minutes ago                        
nja5m0crefrv         \_ wulaoer_web.1   nginx:latest        slave                    Shutdown            Shutdown 2 minutes ago                       
8wazfd9562tj        wulaoer_web.2       nginx:latest        wulaoer.org              Running             Running 9 minutes ago                        
xqyjcwlznjmy        wulaoer_web.3       nginx:latest        node                     Running             Running 2 minutes ago                        
kmntagm4lgk0         \_ wulaoer_web.3   nginx:latest        slave                    Shutdown            Shutdown 2 minutes ago                       
pd8w2f9ivivk        wulaoer_web.4       nginx:latest        node                     Running             Running 9 minutes ago                        
nsld1eq9qoka        wulaoer_web.5       nginx:latest        node                     Running             Running 9 minutes ago                        

节点恢复后不在还原到原来的节点上,而且shutdown状态的应用不能删除,扩展后会继续启用。

应用扩展和伸缩

如果节点不够使用,需要手动增加副本(副本的意思就是容器),使用scale参数,直接在应用名后加扩展后的总数即可。

[root@wulaoer.org  ~]# docker service scale wulaoer_web=10
wulaoer_web scaled to 10
overall progress: 10 out of 10 tasks 
1/10: running   [==================================================>] 
2/10: running   [==================================================>] 
3/10: running   [==================================================>] 
4/10: running   [==================================================>] 
5/10: running   [==================================================>] 
6/10: running   [==================================================>] 
7/10: running   [==================================================>] 
8/10: running   [==================================================>] 
9/10: running   [==================================================>] 
10/10: running   [==================================================>] 
verify: Service converged 
[root@wulaoer.org  ~]# docker service ps wulaoer_web
ID                  NAME                IMAGE               NODE                    DESIRED STATE       CURRENT STATE                ERROR               PORTS
q98hcjj9id87        wulaoer_web.1       nginx:latest        wulaoer.org              Running             Running 7 minutes ago                            
nja5m0crefrv         \_ wulaoer_web.1   nginx:latest        slave                    Shutdown            Shutdown 7 minutes ago                           
8wazfd9562tj        wulaoer_web.2       nginx:latest        wulaoer.org              Running             Running 13 minutes ago                           
xqyjcwlznjmy        wulaoer_web.3       nginx:latest        node                     Running             Running 7 minutes ago                            
kmntagm4lgk0         \_ wulaoer_web.3   nginx:latest        slave                    Shutdown            Shutdown 7 minutes ago                           
pd8w2f9ivivk        wulaoer_web.4       nginx:latest        node                     Running             Running 13 minutes ago                           
nsld1eq9qoka        wulaoer_web.5       nginx:latest        node                     Running             Running 13 minutes ago                           
dz9jcfbapp2b        wulaoer_web.6       nginx:latest        slave                    Running             Running about a minute ago                       
s862hh31sbmm        wulaoer_web.7       nginx:latest        slave                    Running             Running about a minute ago                       
w5pa73p4uegz        wulaoer_web.8       nginx:latest        wulaoer.org              Running             Running about a minute ago                       
dtldonliiy72        wulaoer_web.9       nginx:latest        node                     Running             Running about a minute ago                       
prq1yzvkalp2        wulaoer_web.10      nginx:latest        slave                    Running             Running about a minute ago                       

这时你你会发现原来恢复的节点又开始工作了,我们这个操作在针对集群的节点检查,如果slave这个节点的docker关闭了(也可以理解为服务挂了),应用是否会自动的转移,看下面的测试结果:

[root@slave ~]# systemctl stop docker
[root@wulaoer.org  ~]# docker service ps wulaoer_web
ID                  NAME                 IMAGE               NODE                    DESIRED STATE       CURRENT STATE             ERROR               PORTS
q98hcjj9id87        wulaoer_web.1        nginx:latest        wulaoer.org              Running             Running 11 minutes ago                        
nja5m0crefrv         \_ wulaoer_web.1    nginx:latest        slave                    Shutdown            Shutdown 11 minutes ago                       
8wazfd9562tj        wulaoer_web.2        nginx:latest        wulaoer.org              Running             Running 18 minutes ago                        
xqyjcwlznjmy        wulaoer_web.3        nginx:latest        node                     Running             Running 11 minutes ago                        
kmntagm4lgk0         \_ wulaoer_web.3    nginx:latest        slave                    Shutdown            Shutdown 11 minutes ago                       
pd8w2f9ivivk        wulaoer_web.4        nginx:latest        node                     Running             Running 18 minutes ago                        
nsld1eq9qoka        wulaoer_web.5        nginx:latest        node                     Running             Running 18 minutes ago                        
6zwwgfexw40j        wulaoer_web.6        nginx:latest        node                     Running             Running 7 seconds ago                         
dz9jcfbapp2b         \_ wulaoer_web.6    nginx:latest        slave                    Shutdown            Running 25 seconds ago                        
pke6g8pv0mqq        wulaoer_web.7        nginx:latest        wulaoer.org              Running             Running 7 seconds ago                         
s862hh31sbmm         \_ wulaoer_web.7    nginx:latest        slave                    Shutdown            Running 5 minutes ago                         
w5pa73p4uegz        wulaoer_web.8        nginx:latest        wulaoer.org              Running             Running 5 minutes ago                         
dtldonliiy72        wulaoer_web.9        nginx:latest        node                     Running             Running 5 minutes ago                         
uv05g749zsmm        wulaoer_web.10       nginx:latest        wulaoer.org              Running             Running 7 seconds ago                         
prq1yzvkalp2         \_ wulaoer_web.10   nginx:latest        slave                    Shutdown            Running 25 seconds ago                        

所有应用已经全部转移到正常的节点上了,我们把关闭docker服务的节点在启动一下docker,不过应用并没有恢复到slave节点上,这说明节点只有在改动的时候才会发生变化。下面我们做个测试,在缩减节点的同时,使用一个for循环一直访问某一个节点的80端口,测试一下swarm集群中节点的伸缩对应用是否有影响。

[root@slave ~]# for i in `seq 1000000`;do curl -I -m 10 -o /dev/null -s -w %{http_code} 10.211.55.143;done
[root@wulaoer.org  ~]# docker service scale wulaoer_web=5

查看slave节点一直是200的状态,所以证明节点的伸缩不会影响到应用的使用。

应用管理

我们在swarm集群中添加了很多应用,时间太长了。忘记以前添加了哪些应用怎么查看?添加的应用下线了,如何删除掉?看下面的如何操作。

[root@wulaoer.org ~]# docker service ls
ID                  NAME                MODE                REPLICAS            IMAGE                        PORTS
i6r7hxzcge25        portainer           replicated          0/1                 portainer/portainer:latest   *:9000->9000/tcp
stk7qtdb16zp        tomcat_web          replicated          1/1                 tomcat:latest                *:8080->8080/tcp
r65fxe69rst5        wulaoer_web         replicated          5/5                 nginx:latest                 *:80->80/tcp
[root@wulaoer.org ~]# docker service rm i6r7hxzcge25
i6r7hxzcge25

这里的查看涉及到应用的数量,以及应用的版本和端口,不过这里没有一个自动扩容,我们可以使用脚本代替,完全可以在小环境中使用了。如果发布系统直接更新容器版本即可,我们把代码打包放到容器中,直接导出生成镜像然后在swarm集群中拉取镜像文件进行更新,这里的更新是先下一个更新完上线,一个一个的循环,知道全部上线完成。

[root@wulaoer.org ~]# docker service update --image nginx:1.13 wulaoer_web
wulaoer_web
overall progress: 5 out of 5 tasks 
1/5: running   [==================================================>] 
2/5: running   [==================================================>] 
3/5: running   [==================================================>] 
4/5: running   [==================================================>] 
5/5: running   [==================================================>] 
verify: Service converged 
[root@wulaoer.org ~]# docker service ls
ID                  NAME                MODE                REPLICAS            IMAGE                        PORTS
wqsz7akbjq66        portainer           replicated          0/1                 portainer/portainer:latest   *:9000->9000/tcp
stk7qtdb16zp        tomcat_web          replicated          1/1                 tomcat:latest                *:8080->8080/tcp
r65fxe69rst5        wulaoer_web         replicated          5/5                 nginx:1.13                   *:80->80/tcp

上线的过程不影响用户的使用,整个环境不需要其他的操作,自动完成。还有一种更新是创建新的应用和原应用并存,创建成功后删除旧的应用,这样的好处是在发布的过程中如果有出现问题会自动的回滚,而不需要用户接入,错误率降低很多。

[root@wulaoer.org ~]# docker service update wulaoer_web --replicas 10 --rollback-parallelism 2 --rollback-monitor 20s --rollback-max-failure-ratio .2 --image nginx:1.13 
image nginx:1.13 could not be accessed on a registry to record
its digest. Each node will access nginx:1.13 independently,
possibly leading to different nodes running different
versions of the image.

wulaoer_web
overall progress: 7 out of 10 tasks 
1/10: running   [==================================================>] 
2/10: running   [==================================================>] 
3/10: running   [==================================================>] 
4/10: running   [==================================================>] 
5/10: running   [==================================================>] 
6/10: running   [==================================================>] 
7/10: running   [==================================================>] 
8/10: preparing [=================================>                 ] 
9/10:   
10/10:   
service update paused: update paused due to failure or early termination of task ewcfow914q1s3r1tvttjpt1gb
[root@wulaoer.org ~]# docker service ls
ID                  NAME                MODE                REPLICAS            IMAGE                        PORTS
mws14uzcgg92        portainer           replicated          0/1                 portainer/portainer:latest   *:9000->9000/tcp
stk7qtdb16zp        tomcat_web          replicated          1/1                 tomcat:latest                *:8080->8080/tcp
r65fxe69rst5        wulaoer_web         replicated          10/10               nginx:1.13                   *:80->80/tcp
[root@wulaoer.org ~]# docker service ps wulaoer_web | grep "Running"
k21ghvlr1ory        wulaoer_web.1        nginx:1.13          wulaoer.org                Running             Running 21 minutes ago                                          
sk2qyg8t0zbp        wulaoer_web.2        nginx:1.13          wulaoer.org                Running             Running 21 minutes ago                                          
xk4pcanzbj3x        wulaoer_web.3        nginx:latest        node                       Running             Running less than a second ago                                  
gasdzux6x98s        wulaoer_web.4        nginx:latest        slave                      Running             Running less than a second ago                                  
jfksh7xf6e8k        wulaoer_web.5        nginx:1.13          wulaoer.org                Running             Running 43 seconds ago                                          
jtmcvac12ear        wulaoer_web.6        nginx:1.13          wulaoer.org                Running             Running 21 minutes ago                                          
qrz876zaklez        wulaoer_web.7        nginx:1.13          wulaoer.org                Running             Running 21 minutes ago                                          
5c2wyx2ak3hz        wulaoer_web.8        nginx:1.13          wulaoer.org                Running             Running 21 minutes ago                                          
9m5ztgfqacvi        wulaoer_web.9        nginx:1.13          wulaoer.org                Running             Running 21 minutes ago                                          
ojk6rk5hh5au        wulaoer_web.10       nginx:1.13          wulaoer.org                Running             Running 21 minutes ago                                          

上面的参数注解update更新的服务名称是wulaoer_web。增加了副本是--replicas数量是10。同时执行回滚数动作--rollback-parallelism,数量是2。每次回滚间隔设置--rollback-monitor,时间为2s。回滚的故障率设置--rollback-max-failure-ratio如果小于.2就是百分之0.2允许运行。更新的镜像--image,镜像名称是nginx:1.13。

上面的更新没有问题,那就直接上线,如果更新出现问题我们需要进行回滚,那么回滚的操作设置可以参考下面的信息:

[root@wulaoer.org ~]# docker service update --rollback wulaoer_web
wulaoer_web
rollback: manually requested rollback 
overall progress: rolling back update: 5 out of 5 tasks 
1/5: running   [>                                                  ] 
2/5: running   [>                                                  ] 
3/5: running   [>                                                  ] 
4/5: running   [>                                                  ] 
5/5: running   [>                                                  ] 
verify: Service converged 
[root@wulaoer.org ~]# docker service ls
ID                  NAME                MODE                REPLICAS            IMAGE                        PORTS
wqsz7akbjq66        portainer           replicated          0/1                 portainer/portainer:latest   *:9000->9000/tcp
stk7qtdb16zp        tomcat_web          replicated          1/1                 tomcat:latest                *:8080->8080/tcp
r65fxe69rst5        wulaoer_web         replicated          5/5                 nginx:latest                 *:80->80/tcp

我们的swarm集群已经使用的差不多了,目前只是学习,所以运行的环境不太一样,如果我们使用不同的应用,所需要的CPU和内存是不一样的,所以在创建应用时,需要根据不同的配置然后在node节点上运行不同的应用。下面我们尝试一下在某一个节点上增加应用。swarm集群如果想在指定的节点上运行指定的服务,需要给指定的节点大上标签,打完标签,在发布应用时会自动的跑到指定的节点上。

[root@wulaoer.org ~]# docker node update --label-add func=nginx node     #给节点打标签
node
[root@wulaoer.org ~]# docker service create --name my_nginx --constraint 'node.labels.func == nginx' nginx   #根据标签创建应用
gm51gy4hidv8vgm9yfr3co4c3
overall progress: 1 out of 1 tasks 
1/1: running   [==================================================>] 
[root@wulaoer.org ~]# docker service ps my_nginx
ID                  NAME                IMAGE               NODE                DESIRED STATE       CURRENT STATE                    ERROR               PORTS
axoi4v80rdln        my_nginx.1          nginx:latest        node                 Running             Running less than a second ago                       

如果这个服务不想使用了,我们需要下线,首先把容器数量减少,我们缩减应用。保证所有应用都不在承担用户的请求,然后直接删除应用即可。

[root@wulaoer.org ~]# docker service ls
ID                  NAME                MODE                REPLICAS            IMAGE                        PORTS
mws14uzcgg92        portainer           replicated          0/1                 portainer/portainer:latest   *:9000->9000/tcp
stk7qtdb16zp        tomcat_web          replicated          1/1                 tomcat:latest                *:8080->8080/tcp
r65fxe69rst5        wulaoer_web         replicated          10/10               nginx:1.13                   *:80->80/tcp
[root@wulaoer.org ~]# docker service scale wulaoer_web=0
wulaoer_web scaled to 0
overall progress: 0 out of 0 tasks 
verify: Service converged 
[root@wulaoer.org ~]# docker service ls
ID                  NAME                MODE                REPLICAS            IMAGE                        PORTS
mws14uzcgg92        portainer           replicated          0/1                 portainer/portainer:latest   *:9000->9000/tcp
stk7qtdb16zp        tomcat_web          replicated          1/1                 tomcat:latest                *:8080->8080/tcp
r65fxe69rst5        wulaoer_web         replicated          0/0                 nginx:1.13                   *:80->80/tcp
[root@wulaoer.org ~]# docker service rm wulaoer_web
wulaoer_web
[root@wulaoer.org ~]# docker service ls
ID                  NAME                MODE                REPLICAS            IMAGE                        PORTS
mws14uzcgg92        portainer           replicated          0/1                 portainer/portainer:latest   *:9000->9000/tcp
stk7qtdb16zp        tomcat_web          replicated          1/1                 tomcat:latest                *:8080->8080/tcp

整个应用已经删除完毕,端口也不在占用,可以部署新的应用,不影响新的应用使用。

页面管理swarm

swarm集群也和Kubernetes一样,也有页面管理操作,我们先创建一个swarm集群的页面管理,然后看一下页面管理是什么样子的。

[root@server09 ~]# docker service create  --name portainer --publish 9000:9000 --replicas=1 --constraint 'node.role == manager' --mount type=bind,source=/path/on/host/data,dst=/data portainer/portainer
mp4r44jctqmol9gs06zvwd8g3
overall progress: 1 out of 1 tasks 
1/1: running   [==================================================>] 
verify: Service converged

swarm集群管理页面portainer已经做好了,下面使用浏览器,访问Leader的IP加端口9000,查看一下:

默认没有用户,需要自己创建,输入管理员admin的密码即可,第二次是验证。

登录后,连接一下Leader,我这个是浏览器自动翻译了,应该是英文的。

我这这里把所有的节点都加上去了,加入方法在这里添加。

页面自动获取swarm集群中创建的服务以及副本的情况。还有网络,更新以及回滚的操作。目前真对swarm集群的操作部分就先写到这了,swarm的页面还是可以的,不过对比一下k8s,那就不太好了。

swarm和k8s相比有很多不同点点,但是在使用上差别不是很大,如果运行的服务流量比较稳定的话可以尝试使用swarm集群,不过国内使用的比较少,可能是先入为主吧,很多人都会舍弃swarm而使用k8s。

avatar

发表评论

:?: :razz: :sad: :evil: :!: :smile: :oops: :grin: :eek: :shock: :???: :cool: :lol: :mad: :twisted: :roll: :wink: :idea: :arrow: :neutral: :cry: :mrgreen: