Docker Swarm 服务版本更新
环境：
- 系统：Centos 7.4 x64
- 应用版本：Docker 18.09.0
- 管理节点：192.168.1.79
- 工作节点：192.168.1.78
- 工作节点：192.168.1.77
1、管理节点：创建服务设置更新策略
docker service create \ --replicas 3 \ --name redis \ --update-delay 10s \ --update-parallelism 2 \ --update-failure-action continue \ redis:3.0.6
# 创建命令 docker service create \ # 副本数 --replicas 3 \ # 服务名 --name redis \ # 设定容器间更新时间间隔 --update-delay 10s \ # 更新时同时并行更新数量，默认1 --update-parallelism 2 \ # 任务容器更新失败时的模式 continue为继续使用 --update-failure-action continue \ # 镜像 redis:3.0.6
命令：docker service ls
ID NAME MODE REPLICAS IMAGE
v76p9bl941bk redis replicated 3/3 redis:3.0.6
命令：docker service ps redis ID NAME IMAGE NODE DESIRED STATE CURRENT STATE ERROR PORTS 2pruajw13gud redis.1 redis:3.0.6 slave02 Running Running 4 minutes ago ma8ait8xcagl redis.2 redis:3.0.6 slave02 Running Running 4 minutes ago igpwcuyumwzu redis.3 redis:3.0.6 slave01 Running Running 5 minutes ago
2、管理节点：手动更新服务版本
docker service update --image redis:3.0.7 redis
命令：docker service ps redis ID NAME IMAGE NODE DESIRED STATE CURRENT STATE ERROR PORTS tio9f89x7fzk redis.1 redis:3.0.7 slave02 Running Running 2 minutes ago 2pruajw13gud \_ redis.1 redis:3.0.6 slave02 Shutdown Shutdown 2 minutes ago e3fuv9fm202x redis.2 redis:3.0.7 slave01 Running Running 2 minutes ago ma8ait8xcagl \_ redis.2 redis:3.0.6 slave02 Shutdown Shutdown 2 minutes ago zku5h3h4dqkn redis.3 redis:3.0.7 slave01 Running Running 2 minutes ago igpwcuyumwzu \_ redis.3 redis:3.0.6 slave01 Shutdown Shutdown 3 minutes ago
docker service inspect redis [ { "ID": "v76p9bl941bk87pwwk3nvtwfe", "Version": { "Index": 1479 }, "CreatedAt": "2018-11-07T06:19:23.210686656Z", "UpdatedAt": "2018-11-07T06:30:53.565888582Z", "Spec": { # 服务名称 "Name": "redis", "TaskTemplate": { "ContainerSpec": { # 服务镜像 "Image": "redis:3.0.7@sha256:730b765df9fe96af414da64a2b67f3a5f70b8fd13a31e5096fee4807ed802e20", "DNSConfig": {} }, "Resources": { "Limits": {}, "Reservations": {} }, "RestartPolicy": { "Condition": "any", "MaxAttempts": 0 }, "Placement": {}, "ForceUpdate": 0 }, "Mode": { # 模式 "Replicated": { # 副本数 "Replicas": 3 } }, # 服务更新时应用策略 "UpdateConfig": { # 同时更新任务数 "Parallelism": 1, # 更新间隔时间 "Delay": 10000000000, # 更新失败后动作 pause暂停 "FailureAction": "pause", # 更新失败后继续运行半分比 "MaxFailureRatio": 0 }, "EndpointSpec": { # 默认分配1个虚拟IP "Mode": "vip" } }, "PreviousSpec": { "Name": "redis", "TaskTemplate": { "ContainerSpec": { "Image": "redis:3.0.6@sha256:6a692a76c2081888b589e26e6ec835743119fe453d67ecf03df7de5b73d69842", "DNSConfig": {} }, "Resources": { "Limits": {}, "Reservations": {} }, "RestartPolicy": { "Condition": "any", "MaxAttempts": 0 }, "Placement": {}, "ForceUpdate": 0 }, "Mode": { "Replicated": { "Replicas": 3 } }, "UpdateConfig": { "Parallelism": 1, "Delay": 10000000000, "FailureAction": "pause", "MaxFailureRatio": 0 }, "EndpointSpec": { "Mode": "vip" } }, "Endpoint": { "Spec": {} }, "UpdateStatus": { "State": "completed", "StartedAt": "2018-11-07T06:29:00.44192283Z", "CompletedAt": "2018-11-07T06:30:53.56587273Z", "Message": "update completed" } } ]
注：升级过程，按顺序停止一个容器任务去对这个停止的任务进行更新，最后更新为running状态，再继续往下更新其他容器任务，如果更新失败它会暂停更新，切换到下一个容器任务继续更新直到所有容器。
Docker Swarm 服务版本回滚
1、管理节点：创建服务时设定回滚策略
docker service create \ --name my_web \ --replicas 10 \ --rollback-parallelism 2 \ --rollback-monitor 20s \ --rollback-max-failure-ratio .2 \ nginx:1.12
# 创建服务 docker service create \ # 服务名 --name my_web \ # 副本数 --replicas 10 \ # 同时执行回滚数动作 --rollback-parallelism 2 \ # 每次容器与容器之间的回滚时间间隔 --rollback-monitor 20s \ # 回滚故障率如果小于百分比允许运行 --rollback-max-failure-ratio .2 \ # 使用镜像 nginx:1.12
命令：docker service ps -f "DESIRED-STATE=running" my_web ID NAME IMAGE NODE DESIRED STATE CURRENT STATE ERROR PORTS bgqpm1s0a24l my_web.1 nginx:1.12 slave01 Running Running about a minute ago jttvdgzpiz02 my_web.2 nginx:1.12 slave01 Running Running about a minute ago t8yqfrzoa7fj my_web.3 nginx:1.12 slave02 Running Running about a minute ago xgh712gi2eld my_web.4 nginx:1.12 slave02 Running Running about a minute ago 2tnw3tnofyn5 my_web.5 nginx:1.12 slave01 Running Running about a minute ago on4khcpkb2h7 my_web.6 nginx:1.12 slave02 Running Running about a minute ago gxdeqfbmq60k my_web.7 nginx:1.12 slave01 Running Running about a minute ago kmyz0btgb0hk my_web.8 nginx:1.12 slave02 Running Running about a minute ago lso2ry7f2v0o my_web.9 nginx:1.12 slave01 Running Running 2 minutes ago ee21b2ud0mkk my_web.10 nginx:1.12 slave02 Running Running about a minute ago
2、管理节点：更新服务提升一个版本
docker service update --image nginx:1.13 my_web
docker service ps -f "DESIRED-STATE=running" my_web ID NAME IMAGE NODE DESIRED STATE CURRENT STATE ERROR PORTS bgqpm1s0a24l my_web.1 nginx:1.13 slave01 Running Running about a minute ago jttvdgzpiz02 my_web.2 nginx:1.13 slave01 Running Running about a minute ago t8yqfrzoa7fj my_web.3 nginx:1.13 slave02 Running Running about a minute ago xgh712gi2eld my_web.4 nginx:1.13 slave02 Running Running about a minute ago 2tnw3tnofyn5 my_web.5 nginx:1.13 slave01 Running Running about a minute ago on4khcpkb2h7 my_web.6 nginx:1.13 slave02 Running Running about a minute ago gxdeqfbmq60k my_web.7 nginx:1.13 slave01 Running Running about a minute ago kmyz0btgb0hk my_web.8 nginx:1.13 slave02 Running Running about a minute ago lso2ry7f2v0o my_web.9 nginx:1.13 slave01 Running Running 2 minutes ago ee21b2ud0mkk my_web.10 nginx:1.13 slave02 Running Running about a minute ago
3、管理节点：手动回滚到上一个版本
docker service update --rollback my_web
命令：docker service ls
ID NAME MODE REPLICAS IMAGE PORTS
nikder7y06o2 my_web replicated 10/10 nginx:1.12
本文章为转载内容，我们尊重原作者对文章享有的著作权。如有内容错误或侵权问题，欢迎原作者联系我们进行内容更正或删除文章。
