docker深入2-API示例
2018/1/19
一、目的
1、前言
演示 http API 如何使用，作为入门指引，或许能帮助到人，因为 docker 的 API 文档是自动生成的（看起来是用的 Swagger 有待后续深入研究），我比较笨，没找到太多有用的示例，在自己探索的路上毕竟浪费了点时间，本文若能节约你的时间，是我的幸运。
首先，当然是建议先看官方文档 Docker Engine API and SDKs
https://docs.docker.com/engine/api/
有3个类型的示例：go, python, curl
如果熟悉 go 或者 python 的童鞋，要使用 SDK 来达到自己的目的，建议结合 API 文档看 json output 的字段是否有你需要的，然后去获取它。
https://docs.docker.com/engine/api/v1.30
go 的示例：
https://github.com/opera443399/ops/blob/master/doc/Go/src/abc.com/demo/demoDockerAPI/app.go
题外话：如果要完成任务，其实可以考虑最简单的 API 调用，即利用 docker client 来做（而不是自己去写 shell/go/python 来达到目的）：
~] docker -H your_remote_docker_node_ip version
注1：本次实例是在 docker swarm mode 下使用的，目的是：更新指定服务的镜像。
注2：要在 swarm manager node 上执行。
2、版本
~]# docker version
Client:
Version: 17.06.0-ce
API version: 1.30
Go version: go1.8.3
Git commit: 02c1d87
Built: Fri Jun 23 21:20:36 2017
OS/Arch: linux/amd64
Server:
Version: 17.06.0-ce
API version: 1.30 (minimum version 1.12)
Go version: go1.8.3
Git commit: 02c1d87
Built: Fri Jun 23 21:21:56 2017
OS/Arch: linux/amd64
Experimental: false
二、实例
1、创建一个服务
docker service create --name t001 --publish 22222:80 --detach=true opera443399/whoami:0.9
2、更新服务
1) 目标
service_image_latest="opera443399/whoami:0.7"
service_name="t001"
2) 获取当前服务的版本
service_version_index=$(curl -s -g \
--unix-socket /var/run/docker.sock \
http:/v1.30/services?filters='{"name":["'${service_name}'"]}' \
|jq '.[].Version.Index')
3) 执行更新
curl -s \
--unix-socket /var/run/docker.sock \
"http:/v1.30/services/${service_name}/update?version=${service_version_index}" \
-X POST \
-H "Content-Type: application/json" \
-d "
{
\"Name\": \"${service_name}\",
\"TaskTemplate\": {
\"ContainerSpec\": {
\"Image\": \"${service_image_latest}\"
}
}
}
" |jq '.'
4) 查看服务现状
curl -s -g \
--unix-socket /var/run/docker.sock \
http:/services?filters='{"name":["'${service_name}'"]}' \
|jq '.'
三、问题
1、如果创建 service 时，使用自定义的网络，怎么办？
状态：未解决
1）如果直接使用上述方案，则报错如下
{
"message": "rpc error: code = 12 desc = networks must be migrated to TaskSpec before being changed"
}
2）注意下述关于 Networks 的注释
引用自：https://github.com/moby/moby/blob/master/api/types/swarm/service.go
// ServiceSpec represents the spec of a service.
type ServiceSpec struct {
Annotations
// TaskTemplate defines how the service should construct new tasks when
// orchestrating this service.
TaskTemplate TaskSpec `json:",omitempty"`
Mode ServiceMode `json:",omitempty"`
UpdateConfig *UpdateConfig `json:",omitempty"`
RollbackConfig *UpdateConfig `json:",omitempty"`
// Networks field in ServiceSpec is deprecated. The
// same field in TaskSpec should be used instead.
// This field will be removed in a future release.
Networks []NetworkAttachmentConfig `json:",omitempty"`
EndpointSpec *EndpointSpec `json:",omitempty"`
}
引用自：https://github.com/moby/moby/blob/master/api/types/swarm/task.go
// TaskSpec represents the spec of a task.
type TaskSpec struct {
// ContainerSpec and PluginSpec are mutually exclusive.
// PluginSpec will only be used when the `Runtime` field is set to `plugin`
ContainerSpec *ContainerSpec `json:",omitempty"`
PluginSpec *runtime.PluginSpec `json:",omitempty"`
Resources *ResourceRequirements `json:",omitempty"`
RestartPolicy *RestartPolicy `json:",omitempty"`
Placement *Placement `json:",omitempty"`
Networks []NetworkAttachmentConfig `json:",omitempty"`
// LogDriver specifies the LogDriver to use for tasks created from this
// spec. If not present, the one on cluster default on swarm.Spec will be
// used, finally falling back to the engine default if not specified.
LogDriver *Driver `json:",omitempty"`
// ForceUpdate is a counter that triggers an update even if no relevant
// parameters have been changed.
ForceUpdate uint64
Runtime RuntimeType `json:",omitempty"`
}
3）尝试过多种在 API 中增加网络相关的配置，均未找到合理的方法
示例：
TaskTemplate.Networks
{
\"Name\": \"${service_name}\",
\"TaskTemplate\": {
\"ContainerSpec\": {
\"Image\": \"${service_image_latest}\"
},
\"Networks\": [
{
\"Target\": \"xxx\"
}
]
}
}
注：执行有风险，效果是：
ingress网络消失，该 service 对外发布的端口消失。
因为，创建 service 时：
docker service create --name t001 --network t001only --publish 22222:80 --detach=true opera443399/whoami:0.7
使用 --network 将关联到一个网络 t001only
使用 --publish 将关联到一个网络 ingress
因此，实际上有2个网络。
注1：在反复测试的过程中，出现一个奇怪的现象，，创建 service 时，，容器处于 new 的状态，无法上线，暂时还未找到原因，因而中止了测试。
注2：或许可以尝试在更新 service 的过程中，指定 ip 和 port 等信息，待后续测试后再更新本段信息。
相关报错：
~]# journalctl -u docker -a -S "2017-09-18 14:00" |sed -r "s/(.*)msg=(.*)/\2/g" |grep rpc |sort |uniq
"Error updating service t001: rpc error: code = 12 desc = networks must be migrated to TaskSpec before being changed"
"Error creating service t001: rpc error: code = 4 desc = context deadline exceeded"
"Error removing service t001: rpc error: code = 4 desc = context deadline exceeded"
"Error updating service t001: rpc error: code = 2 desc = update out of sequence"
"Handler for POST /v1.30/services/t001/update returned error: rpc error: code = 12 desc = networks must be migrated to TaskSpec before being changed"
"Handler for POST /services/create returned error: rpc error: code = 4 desc = context deadline exceeded"
"Handler for DELETE /v1.30/services/t001 returned error: rpc error: code = 4 desc = context deadline exceeded"
"Handler for POST /v1.30/services/t001/update returned error: rpc error: code = 2 desc = update out of sequence"
2、昨天的测试后，出现创建的 service 无法正常运行的错误追踪
从头开始，继续排错。
1）创建一个 global 类型的服务
~]# docker service create --name t002 --publish 10008:80 --mode global --detach=true nginx:latest
2）查看服务状态
~]# docker service ls -f 'name=t002'
ID NAME MODE REPLICAS IMAGE PORTS
5a8nps5pssho t002 global 0/4 nginx:latest *:10008->80/tcp
~]# docker service ps t002
ID NAME IMAGE NODE DESIRED STATE CURRENT STATE ERROR PORTS
4vr4ehhj17l3 t002.node1_id_here nginx:latest node1.test.com Running New 21 minutes ago
0k2ttjveis6t t002.node2_id_here nginx:latest node2.test.com Running New 21 minutes ago
wwopgjy8pfxj t002.node3_id_here nginx:latest node3.test.com Running New 21 minutes ago
果然，没有端口映射，服务还没上线：
~]# docker service inspect --pretty t002
ID: 5a8nps5psshoemixl8w6iygss
Name: t002
Service Mode: Global
Placement:
UpdateConfig:
Parallelism: 1
On failure: pause
Monitoring Period: 5s
Max failure ratio: 0
Update order: stop-first
RollbackConfig:
Parallelism: 1
On failure: pause
Monitoring Period: 5s
Max failure ratio: 0
Rollback order: stop-first
ContainerSpec:
Image: nginx:latest
Resources:
Endpoint Mode: vip
查看 task 信息：
curl -s --unix-socket /var/run/docker.sock http:/tasks |jq '.' >tmp.1
{
"ID": "wwopgjy8pfxj4baa3jrn438rm",
"Version": {
"Index": 5070352
},
"CreatedAt": "2017-09-19T03:22:07.779333332Z",
"UpdatedAt": "2017-09-19T03:22:07.779333332Z",
"Labels": {},
"Spec": {
"ContainerSpec": {
"Image": "nginx:latest"
},
"Placement": {},
"ForceUpdate": 0
},
"ServiceID": "5a8nps5psshoemixl8w6iygss",
"NodeID": "node2_id_here",
"Status": {
"Timestamp": "2017-09-19T03:22:07.77932103Z",
"State": "new",
"Message": "created",
"ContainerStatus": {},
"PortStatus": {}
},
"DesiredState": "running"
}
该 task 的状态不符合预期，最终 service 启动失败。
3）日志
~]# journalctl -u docker -a -S "2017-09-19 11:00" |sed -r "s/(.*)msg=(.*)/\2/g" |sort |uniq |grep -vE "(Logs begin|taskmanager|no pull access|Node join event|for pull|authentication|raft)"
"Could not parse VIP address while releasing"
"error deallocating vip" error="invalid CIDR address: " vip.addr= vip.network=fjvkitgqzewmtujakidzo38pk
"Failed allocation for service 5a8nps5psshoemixl8w6iygss" error="could not find an available IP while allocating VIP" module=node node.id=node2_id_here
"task allocation failure" error="service 5a8nps5psshoemixl8w6iygss to which this task 0k2ttjveis6tliuxqyuxqyp97 belongs has pending allocations" module=node node.id=node2_id_here
"task allocation failure" error="service 5a8nps5psshoemixl8w6iygss to which this task 4vr4ehhj17l3f768ksc14w6rz belongs has pending allocations" module=node node.id=node2_id_here
"task allocation failure" error="service 5a8nps5psshoemixl8w6iygss to which this task wwopgjy8pfxj4baa3jrn438rm belongs has pending allocations" module=node node.id=node2_id_here
初步看起来，似乎是网络方面的问题，没有可用的 VIP 了？神奇，，待查证。
截止到目前，并未找到原因，为了节约时间，只能采取重启 docker 服务后，恢复集群到正常的状态，最后调用前文提及的 docker client 通过远程 API 来执行的方式完成任务先，后续有空再尝试重放现场。
ZYXW、参考
1、API
https://docs.docker.com/engine/api/
https://docs.docker.com/engine/api/v1.30
2、moby src
https://github.com/moby/moby/blob/master/api/types/swarm/task.go
https://github.com/moby/moby/blob/master/api/types/swarm/service.go
3、portainer src
https://github.com/portainer/portainer/blob/04ea81e7cd8401690058c4b4264452bf9d7a05eb/app/components/service/serviceController.js
转载于:https://blog.51cto.com/nosmoking/1966417
扫一扫
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
