模块下载 & 同步
git submodule update --init --recursive安装依赖
pnpm install以 Dev 模式转译所有包
pnpm -r dev配置 ft 命令行工具
命令行工具入口位于 packages/faasit/bin/faasit.cjs
命令行工具脚本位于 scripts/ft
为了方便开发，可将脚本 scripts/ft 路径添加到 PATH 变量中
export PATH=$PATH:"<faasit-src-dir>/scripts"之后能够在任意位置使用 ft 命令
运行 ft 命令行工具
cd demo-202405/helloworld/
# 将 DSL 转换为 IR
ft eval main.ft
# 部署函数到云平台
ft deploy -p Aliyun
# 调用部署好的函数
ft invoke -p Aliyun
# 在本地运行 Function & Workflow
ft run
# 执行代码生成
ft codegen
# 新建 Function 项目目录
ft init- Docker
- 网络配置
sudo apt-get install -y selinux-utils
sudo setenforce 0
sudo ufw disable
sudo swapoff -a
sudo modprobe br_netfilter
echo "net.bridge.bridge-nf-call-ip6tables = 1" | sudo tee -a /etc/sysctl.conf
echo "net.bridge.bridge-nf-call-iptables = 1" | sudo tee -a /etc/sysctl.conf
echo "net.ipv4.ip_forward = 1" | sudo tee -a /etc/sysctl.conf- 使用阿里云的镜像构建k8s
sudo apt-get update && sudo apt-get install -y apt-transport-https
curl https://mirrors.aliyun.com/kubernetes/apt/doc/apt-key.gpg | sudo apt-key add -- 将阿里云的镜像地址写到
sources.list中
sudo vim /etc/apt/sources.list.d/kubernetes.list
# 写入下列内容
deb https://mirrors.aliyun.com/kubernetes/apt/ kubernetes-xenial main- 安装特定版本的k8s
sudo apt-get update
sudo apt-get install -y kubelet==1.27.0-00 kubeadm==1.27.0-00 kubectl==1.27.0-00- 安装
cri-dockered插件
wget https://github.com/Mirantis/cri-dockerd/releases/download/v0.3.4/cri-dockerd-0.3.4.amd64.tgz
tar xvf cri-dockerd-0.3.4.amd64.tgz
sudo mv cri-dockerd/cri-dockerd /usr/local/bin/
wget https://raw.githubusercontent.com/Mirantis/cri-dockerd/master/packaging/systemd/cri-docker.service
wget https://raw.githubusercontent.com/Mirantis/cri-dockerd/master/packaging/systemd/cri-docker.socket
sudo mv cri-docker.* /etc/systemd/system/
sudo systemctl daemon-reload
sudo systemctl enable cri-docker.service
sudo systemctl enable --now cri-docker.socketkubeadm-config.yaml配置
apiVersion: kubeadm.k8s.io/v1beta3
bootstrapTokens:
- groups:
- system:bootstrappers:kubeadm:default-node-token
token: abcdef.0123456789abcdef
ttl: 24h0m0s
usages:
- signing
- authentication
kind: InitConfiguration
localAPIEndpoint:
advertiseAddress: 192.168.28.220 # 自行修改
bindPort: 6443
nodeRegistration:
criSocket: /var/run/cri-dockerd.sock
imagePullPolicy: IfNotPresent
name: cloud
taints: null
---
apiServer:
timeoutForControlPlane: 4m0s
apiVersion: kubeadm.k8s.io/v1beta3
certificatesDir: /etc/kubernetes/pki
clusterName: kubernetes
controllerManager: {}
dns: {}
etcd:
local:
dataDir: /var/lib/etcd
imageRepository: registry.aliyuncs.com/google_containers
kind: ClusterConfiguration
kubernetesVersion: 1.27.0
networking:
dnsDomain: cluster.local
serviceSubnet: 10.96.0.0/12
podSubnet: 10.244.0.0/16
scheduler: {}- 启动k8s集群
sudo kubeadm init --config kubeadm-config.yaml
sudo mkdir -p .kube/config
sudo cp -i /etc/kubernetes/admin.conf .kube/config
sudo chown $(id -u):$(id -g) .kube/configkubectl taint node epyc node-role.kubernetes.io/control-plane-
kubectl apply -f https://github.com/knative/serving/releases/download/knative-v1.12.4/serving-crds.yaml
kubectl apply -f https://github.com/knative/serving/releases/download/knative-v1.12.4/serving-core.yaml
kubectl apply -f https://github.com/knative/net-kourier/releases/download/knative-v1.12.3/kourier.yaml
kubectl patch configmap/config-network \
--namespace knative-serving \
--type merge \
--patch '{"data":{"ingress-class":"kourier.ingress.networking.knative.dev"}}'
kubectl apply -f https://github.com/knative/serving/releases/download/knative-v1.12.4/serving-default-domain.yaml
# 10.0.0.233 is an arbitary choice.EXTERNAL_IP="10.0.0.233"
# To get rid of the strange rules that default urls *.svc.cluster.local cannot be accessed from outside network.
# sslip can avoid us from trouble of manipulating DNS record.
export EXTERNAL_IP=10.0.0.233
kubectl patch configmap/config-domain \--namespace knative-serving \--type merge \--patch "{\"data\":{\"$EXTERNAL_IP.sslip.io\":\"\"}}"
kubectl patch svc kourier -n kourier-system -p "{\"spec\": {\"type\": \"LoadBalancer\", \"externalIPs\": [\"$EXTERNAL_IP\"]}}"- 验证外部IP
kubectl --namespace kourier-system get service kourierdocker pull pypiserver/pypiserver:v2.2.0
mkdir -p ~/.pypi
cp deploy_pku/htpasswd.txt ~
docker run --restart always --name pypi -d -p 12121:8080 -v ~/.pypi/:/data/packages -v ~/htpasswd.txt:/data/.htpasswd pypiserver/pypiserver:v2.2.0 run -P .htpasswd packagesdocker pull registry:2
docker run --restart always -d -p 5000:5000 --name registry -v /root/dockerimage:/var/lib/registry registry:2- 上传
faasit-runtime
cd faasit-runtime/faasit-python-runtime
bash build.sh- 构建镜像
cd faasit/faasit-docker/kn
# 可自行修改build.sh中的版本信息
bash build.sh- 通过阿里云账户得到如下信息并写到环境变量中
export FAASIT_SECRET_ALIYUN_ACCOUNT_ID=176xxxxxxxxx
export ALIBABA_CLOUD_OSS_BUCKET_NAME=faasit
export ALIBABA_CLOUD_OSS_REGION=oss-cn-hangzhou
export FAASIT_SECRET_ALIYUN_ACCESS_KEY_ID=LTAxxxxxxxxxxxxxxxx
export FAASIT_SECRET_ALIYUN_ACCESS_KEY_SECRET=o3vxxxxxxxxxxxxxxxxx
export FAASIT_SECRET_ALIYUN_REGION=cn-hangzhou前提
- Python3.10以上的环境
- Docker
方案1：自己构建faasit-runtime
- 搭建私有pypi服务器
docker pull pypiserver/pypiserver:v2.2.0
mkdir -p ~/.pypi
cp deploy_pku/htpasswd.txt ~
docker run --rm --name pypi -d -p 12121:8080 -v ~/.pypi/:/data/packages -v ~/htpasswd.txt:/data/.htpasswd pypiserver/pypiserver:v2.2.0 run -P .htpasswd packages- 构建faasit-runtime并上传(可能需要下载wheel,twine等工具，遇到了自己
pip install即可)
cd faasit-runtime/faasit-python-runtime
bash build.sh- 之后在pip的install路径下添加一下
localhost:12121(自行gpt查一下)
方案2：使用已有的faasit-runtime
其实我们已经在pypi上传了faasit-runtime
cd faasit-docker/spilot/0.6
docker build --no-cache -t faasit-spilot:0.6 如果上面使用faasit-runtime使用的是我们上传的运行时环境，就把pip install那一步的后面的url删掉即可
前提
密钥放在.ssh/demo.pem下
将deploy_pku/python.sh复制到北大的目录下
cd deploy_pku
scp -i ~/.ssh/demo.pem python.sh root@xx.xx.xx.xx:/root登录到北大服务器
bash python.shcd deploy_pku
curl --location --fail https://github.com/volta-cli/volta/releases/download/v2.0.2/volta-2.0.2-linux.tar.gz --output volta-2.0.2-linux.tar.gz
scp -i ~/.ssh/demo.pem volta-2.0.2-linux.tar.gz root@xx.xx.xx.xx:/root
scp -i ~/.ssh/demo.pem volta.sh root@xx.xx.xx.xx:/root登录到北大服务器
bash volta.sh
volta install node@16
volta install pnpm@8.6.0pwd # faasit
cd ..
scp -i ~/.ssh/demo.pem -r faasit root@xx.xx.xx.xx:/root登录到北大服务器
cd faasit
pnpm i
pnpm -r devcd deploy_pku
docker pull registry:2
docker save registry:2 -o registry.tar
scp -i ~/.ssh/demo.pem registry.tar root@xx.xx.xx.xx:/root登录北大服务器master节点
mkdir -p dockerimage
docker load -i registry.tar
docker run --rm -d -p 5000:5000 --name registry -v /root/dockerimage:/var/lib/registry registry:2使用ip addr获取北大master节点的内网IP地址，大概是192.168.0.xxx
然后登录北大服务器的其他node节点
sudo vim /etc/docker/daemon.json写入以下内容
{
"insecure-registries": ["192.168.0.xxx:5000"],
"registry-mirrors": ["http://192.168.0.xxx:5000"]
}之后重新加载一下docker的配置
sudo systemctl reload docker可以通过下列命令测试环境是否部署成功
curl -X GET http://192.168.0.136:5000/v2/_catalog
# {"repositories":["library/redis","redis"]}docker pull pypiserver/pypiserver:v2.2.0
docker save pypiserver/pypiserver:v2.2.0 -o pypi.tar
scp -i ~/.ssh/demo.pem pypi.tar root@xx.xx.xx.xx:/root
scp -i ~/.ssh/demo.pem htpasswd.txt root@xx.xx.xx.xx:/root登录北大服务器
mkdir -p .pypi
docker load -i pypi.tar
docker run --rm --name pypi -d -p 12121:8080 -v ~/.pypi/:/data/packages -v ~/htpasswd.txt:/data/.htpasswd pypiserver/pypiserver:v2.2.0 run -P .htpasswd packages上传serverless-framework以及faasit-runtime
cd /root/faasit/faasit-runtime/faasit-python-runtime/pku-pkg
pip3 install wheel twine
python3 setup.py bdist_wheel
twine upload --repository-url http://localhost:12121/ dist/* -u "faasit" -p "faasit-pypi" --verbose
cd ..
python3 setup.py bdist_wheel
twine upload --repository-url http://localhost:12121/ dist/* -u "faasit" -p "faasit-pypi" --verbose安装的时候只需要
pip install faasit-runtime --index-url http://localhost:12121如果是在其他节点
pip install faasit-runtime --index-url http://{ip}:12121 --trust-host {ip}下载code-server并上传
wget https://github.com/coder/code-server/releases/download/v4.96.1/code-server_4.96.1_amd64.deb
scp -i ~/.ssh/demo.pem code-server_4.96.1_amd64.deb pkualiyun:/coder登录北大服务器
useradd -mG sudo coder
passwd coder
# faasitcoder
visudo
# add below
# coder ALL=(ALL:ALL) NOPASSWD: ALL
chsh coder -s /bin/bash
dpkg -i /home/coder/code-server_4.96.1_amd64.deb
vim /home/coder/.config/code-server/config.yaml
# 写入下列内容
bind-addr: 0.0.0.0:8080
auth: password
password: faasit-code
cert: false
vim /lib/systemd/system/code-server\@.service
# 写入下列内容
[Unit]
Description=code-server
After=network.target
[Service]
Type=exec
ExecStart=/usr/bin/code-server /home/coder/projects
Restart=always
User=%i
[Install]
WantedBy=default.target
systemctl daemon-reload
systemctl restart code-server@coder前置
假设已经写好了应用以及main.ft
@function表示一个函数体，随后跟着函数名称，该函数名称是Serverless应用中对于每个函数的标识符
@function hello {
runtime = "python" # 运行时
codeDir = "./code" # 代码所在的目录
handler = "index.hello" # 响应请求的文件以及对应的函数名称，注意这里指的是Python文件中的名称
}@provider是云平台的一个概念，可以描述不同的部署方式
@provider pku {
kind = 'pku' # 指定部署平台为pku
deployment = {} # 指定部署参数，根据具体情况而定
invoke = {} # 调用参数，对应的`python -m serverless_framework.controller`后跟着的参数
}@application是一个应用的概念
@application helloapp {
functions = [hello] # 这个应用所包含的函数
defaultProvider = pku # 默认部署平台
providers = [pku] # 支持的部署平台
inputExamples = [ # 应用参数
{
value = {
...
}
}
]
}- 开发示例
from faasit_runtime import function, workflow, Workflow, FaasitRuntime # 导入统一运行时环境
@function # 定义一个Serverless函数
def hello(frt: FaasitRuntime): # 统一运行时参数
_in = frt.input() # 函数输入
result = frt.call('hello2', {'a': _in['a']}) # 函数通信，调用另一个函数，其中参数为{'a': _in['a']}
return frt.output(result) # 函数返回
@function
def hello2(frt: FaasitRuntime):
return frt.output({"message":"ok"})
hello = hello.export() # 导出函数到云平台, 这里变量的名称可以自定义，但是要注意跟`main.ft`中的handler定义相对应
hello2 = hello2.export() 构建应用镜像
ft build部署
ft deploy调用
ft invoke
ft invoke --file [] # 指定文件触发报错解决
如果遇到ft deploy的时候输出一些奇怪的东西，则将下列代码写到跟workflow代码的同个目录下并命名为driver.py
import json
import os
os.environ["FAASIT_PROVIDER"]="pku"
from index import handler
output = handler()
print(output)然后运行driver.py根据报错提示修改
-
开发工具
- pnpm
- unbuild
- volta
-
使用 Pnpm 的原因
- 原生支持 Monorepo 项目架构
- 依赖安装速度快
The project is based on https://github.com/cdietrich/xtext-languageserver-example.git