Swarm是Docker官方提供的一款集羣管理工具,其主要做用是把若干臺Docker主機抽象爲一個總體,而且經過一個入口統一管理這些Docker主機上的各類Docker資源。Swarm和Kubernetes比較相似,可是更加輕,具備的功能也較kubernetes更少一些。node
swarm架構圖nginx
多個manager節點就會涉及到狀態的同步,docker swarm採用的是內置的分佈式存儲的數據庫,採用raft協議去同步多個manager之間的狀態,防止腦裂狀況發生。算法
業務通常都會運行在worker節點上(manager也是能夠運行的),多個worker節點之間數據和狀態會經過gossip的網絡同步。docker
docker swarm中,service的概念和docker compose中的service概念是差很少的。shell
replicas是在service的基礎上作橫向擴展,例如,一個nginx的swevice中有三個nginx的容器,其中每個容器就能夠理解爲一個replica。數據庫
swarm中經過manager節點去部署一個service的時候,咱們事先是不知道這個service最終會運行在哪些swarm的cluster節點上的,swarm的scheduler會根據必定的調度算法去計算(好比會根據每臺cluster節點的資源使用率等),將容器調度在相應的節點上。服務器
swarm服務建立和調度圖:網絡
初始化一個集羣,宣告manager節點地址:架構
docker swarm init --advertise-addr=192.168.205.10 Swarm initialized: current node (xyl16kq5aj71sbqccys1jsoyv) is now a manager. To add a worker to this swarm, run the following command: docker swarm join --token SWMTKN-1-5igsd3yoz1hm8z39q0emtgx4ot0ei31vgv8kkvy1jybxwd8286-0lewqgbggwmhrpclt7myjvuu9 192.168.205.10:2377 To add a manager to this swarm, run 'docker swarm join-token manager' and follow the instructions.
到此爲止,docker swarm的manager節點建立完成,能夠經過提示在worker節點上運行下面的命令便可在這個manager節點中添加worker節點:分佈式
docker swarm join --token SWMTKN-1-5igsd3yoz1hm8z39q0emtgx4ot0ei31vgv8kkvy1jybxwd8286-0lewqgbggwmhrpclt7myjvuu9 192.168.205.10:2377
docker swarm join --token SWMTKN-1-5igsd3yoz1hm8z39q0emtgx4ot0ei31vgv8kkvy1jybxwd8286-0lewqgbggwmhrpclt7myjvuu9 192.168.205.10:2377 This node joined a swarm as a worker.
經過docker node ls
查看當前的swarm的節點信息:
docker node ls ID HOSTNAME STATUS AVAILABILITY MANAGER STATUS ENGINE VERSION xyl16kq5aj71sbqccys1jsoyv * swarm-manager Ready Active Leader 18.03.1-ce r2mh5nz1k7d765j6zxj3e1hgf swarm-worker1 Ready Active 18.03.1-ce
docker swarm join --token SWMTKN-1-5igsd3yoz1hm8z39q0emtgx4ot0ei31vgv8kkvy1jybxwd8286-0lewqgbggwmhrpclt7myjvuu9 192.168.205.10:2377 This node joined a swarm as a worker.
docker node ls ID HOSTNAME STATUS AVAILABILITY MANAGER STATUS ENGINE VERSION xyl16kq5aj71sbqccys1jsoyv * swarm-manager Ready Active Leader 18.03.1-ce 81zzel5tk5im0rmp98xi306qa swarm-worker1 Ready Active 18.03.1-ce r2mh5nz1k7d765j6zxj3e1hgf swarm-worker2 Ready Active 18.03.1-ce
經過docker service create
建立一個服務,相似於docker run
命令:
docker service create --name demo busybox sh -c "while true;do sleep 3600;done" i7jn9sbf4xmdv3t6nl8f7ywp3 overall progress: 1 out of 1 tasks 1/1: running [==================================================>] verify: Service converged
查看已有的service:
docker service ls ID NAME MODE REPLICAS IMAGE PORTS i7jn9sbf4xmd demo replicated 1/1 busybox:latest
經過上面的信息咱們不知道demo
這個服務的容器運行在哪一臺服務器上,可是,能夠經過docker service ps service_name
來查看:
docker service ps demo ID NAME IMAGE NODE DESIRED STATE CURRENT STATE ERROR PORTS ewkxqwwa7u5n demo.1 busybox:latest swarm-manager Running Running 4 minutes ago
能夠發現,demo
這個service運行在swarm-manager
這個管理節點上。在swarm-manager
節點上經過docker ps
查看:
docker ps CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES e80a085916b0 busybox:latest "sh -c 'while true;d…" 6 minutes ago Up 6 minutes demo.1.ewkxqwwa7u5ndaaigls3xvts3
咱們經過以前的demo這個service能夠看到有一個REPLICAS
顯示的是1/1
,這就說明demo這個service是能夠橫向擴展的,service的擴展相似於docker compose。也是經過參數scale
來實現:
docker service scale demo=5 demo 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
查看demo
服務:
docker service ls ID NAME MODE REPLICAS IMAGE PORTS i7jn9sbf4xmd demo replicated 5/5 busybox:latest
能夠發現REPLICAS
已經由1/1
變成了5/5
。再次查看這5個容器分別運行在哪些節點之上:
docker service ps demo ID NAME IMAGE NODE DESIRED STATE CURRENT STATE ERROR PORTS ewkxqwwa7u5n demo.1 busybox:latest swarm-manager Running Running 14 minutes ago pxdw4jdn5clq demo.2 busybox:latest swarm-manager Running Running 3 minutes ago 5iayi08grmzo demo.3 busybox:latest swarm-worker1 Running Running 2 minutes ago 313zmkgrm933 demo.4 busybox:latest swarm-worker2 Running Running 3 minutes ago jawwrmdgrey4 demo.5 busybox:latest swarm-worker2 Running Running 3 minutes ago
經過docker swarm scale建立的橫向擴展(好比scale demo=5),那麼,不管在哪一個節點上強制刪除其中一個或者多個該service的容器(dockerd服務必須在運行),docker swarm scale都會自動在當前的集羣中的其餘容器從新啓動相應個數的容器,保持scale指定的數量。這樣就能夠確保系統穩定,不會出現業務中止的狀況。
刪除一個service只須要經過docker service rm service_name
便可,在刪除service以後,docker swarm會自動從當前swarm集羣中刪除該服務以前運行的全部容器(有必定的延時):
docker service rm demo demo
刪除服務以後,再次查看該服務狀況:
docker service ps demo no such service: demo docker ps CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
[root@docker ~]# [root@docker ~]# [root@docker ~]#