冲突paddle-serving-client 0.8.3 depends on grpcio＜=1.33.2 pymilvus 2.0.2 depends on grpcio==1.37.1

问题：paddle-serving和pymilvus包的底层依赖安装包冲突

[2022-05-10 12:06:31] The conflict is caused by:
[2022-05-10 12:06:31]     paddle-serving-server-gpu 0.8.3.post101 depends on grpcio<=1.33.2
[2022-05-10 12:06:31]     paddle-serving-client 0.8.3 depends on grpcio<=1.33.2
[2022-05-10 12:06:31]     pymilvus 2.0.2 depends on grpcio==1.37.1
[2022-05-10 12:06:31] 
[2022-05-10 12:06:31] To fix this you could try to:
[2022-05-10 12:06:31] 1. loosen the range of package versions you've specified
[2022-05-10 12:06:31] 2. remove package versions to allow pip attempt to solve the dependency

解决方案，重新Serving编包

参考文件：https://github.com/PaddlePaddle/Serving/blob/v0.8.3/doc/Compile_CN.md

修改requirements.txt和编译参数中grpc版本依赖，重新Serving编包
修改requirements.txt和 go install google.golang.org/[email protected]，都改成高版本
编译出来的包是0.0.0.0，不需要上传到pypi了，自己用就行

0、准备工作 0.1、下拉包含第三方依赖的镜像

docker pull registry.baidubce.com/paddlepaddle/serving:0.8.0-cuda11.2-cudnn8-devel-with-codes

0.2、创建容器

nvidia-docker run --rm -it registry.baidubce.com/paddlepaddle/serving:0.8.0-cuda11.2-cudnn8-devel-with-codes bash

0.3、拷贝容器的第三方依赖到宿主机

docker cp 6225f86a207e:/home/Serving/server-build-gpu-opencv/third_party/ /data/mart/build_server/third_party/
docker cp 6225f86a207e:/home/Serving/client-build/third_party/ /data/mart/build_client/third_party/
docker cp 6225f86a207e:/root/.cache/ /data/mart/root/.cache/
docker cp 6225f86a207e:/home/Serving/core/cube/cube-agent/src/github.com/ /data/mart/github/

1、编译环境准备，我们首先要针对自己所需的环境拉取相关镜像。上表环境一列下，除了CPU，其余（Cuda**+Cudnn**）都属于GPU环境。 您可以使用Serving开发镜像。

docker pull registry.baidubce.com/paddlepaddle/serving:0.8.0-cuda10.1-cudnn7-devel

# 如果是GPU镜像
nvidia-docker run --rm -it registry.baidubce.com/paddlepaddle/serving:0.8.0-cuda10.1-cudnn7-devel bash

2、下载代码库

子模块，讲解：https://blog.csdn.net/guotianqing/article/details/82391665

git clone https://github.com/PaddlePaddle/Serving
cd Serving && git submodule update --init --recursive

3、环境变量准备 3.1、设置PYTHON环境变量 3.1.1、设置PYTHON_INCLUDE_DIR

find / -name Python.h

export PYTHON_INCLUDE_DIR=/usr/local/include/python3.7m/

3.1.2、设置PYTHON_LIBRARIES

find / -name libpython3.7.so
find / -name libpython3.7m.so

export PYTHON_LIBRARIES=/usr/local/lib

3.1.3、设置PYTHON_EXECUTABLE

which python3.7

export PYTHON_EXECUTABLE=/usr/local/bin/python3.7

4、更改requirements.txt

修改requirements.txt和 go install google.golang.org/[email protected]，都改成高版本

cd python
vim requirements.txt
i
:wq
cd ..

5、配置安装

修改requirements.txt和 go install google.golang.org/[email protected]，都改成高版本

export GOPATH=$HOME/go
export PATH=$PATH:$GOPATH/bin

python3.7 -m pip install -r python/requirements.txt -i https://mirror.baidu.com/pypi/simple

go env -w GO111MODULE=on
go env -w GOPROXY=https://goproxy.cn,direct
go install github.com/grpc-ecosystem/grpc-gateway/protoc-gen-grpc-gateway@v1.15.2
go install github.com/grpc-ecosystem/grpc-gateway/protoc-gen-swagger@v1.15.2
go install github.com/golang/protobuf/protoc-gen-go@v1.4.3
go install google.golang.org/grpc@v1.37.1
go env -w GO111MODULE=auto

6、配置GPU

export CUDA_PATH='/usr/local/cuda'
export CUDNN_LIBRARY='/usr/local/cuda/lib64/'
export CUDA_CUDART_LIBRARY="/usr/local/cuda/lib64/"
export TENSORRT_LIBRARY_PATH="/usr/"

7、编译paddle-serving-server 7.1 把容器的第三方依赖从宿主机拷贝进来

mkdir build_server
mkdir build_client


docker cp /data/mart/build_server/third_party/ ea1301ebf2a3:/home/Serving/build_server/third_party/
docker cp /data/mart/build_client/third_party/ ea1301ebf2a3:/home/Serving/build_client/third_party/
docker cp /data/mart/root/.cache/ ea1301ebf2a3:/root/.cache/ 
docker cp /data/mart/github.com/ ea1301ebf2a3:/home/Serving/core/cube/cube-agent/src/ 

7.2 执行编译

cd build_server
cmake -DPYTHON_INCLUDE_DIR=$PYTHON_INCLUDE_DIR \
    -DPYTHON_LIBRARIES=$PYTHON_LIBRARIES \
    -DPYTHON_EXECUTABLE=$PYTHON_EXECUTABLE \
    -DCUDA_TOOLKIT_ROOT_DIR=${CUDA_PATH} \
    -DCUDNN_LIBRARY=${CUDNN_LIBRARY} \
    -DCUDA_CUDART_LIBRARY=${CUDA_CUDART_LIBRARY} \
    -DTENSORRT_ROOT=${TENSORRT_LIBRARY_PATH} \
    -DSERVER=ON \
    -DWITH_GPU=ON ..
make -j20
cd ..

7.3、编译paddle-serving-client 和 paddle-serving-app

# 编译paddle-serving-client
cd build_client
cmake -DPYTHON_INCLUDE_DIR=$PYTHON_INCLUDE_DIR \
    -DPYTHON_LIBRARIES=$PYTHON_LIBRARIES \
    -DPYTHON_EXECUTABLE=$PYTHON_EXECUTABLE \
    -DCLIENT=ON ..
make -j10
cd ..

# 编译paddle-serving-app
mkdir build_app
cd build_app
cmake -DPYTHON_INCLUDE_DIR=$PYTHON_INCLUDE_DIR \
    -DPYTHON_LIBRARIES=$PYTHON_LIBRARIES \
    -DPYTHON_EXECUTABLE=$PYTHON_EXECUTABLE \
    -DAPP=ON ..
make -j10
cd ..

8、把编译成功的包拷贝到本地

docker cp 8a3eb7d0bde0:/home/Serving/build_server/python/dist/ /data/mart/build_server
docker cp 8a3eb7d0bde0:/home/Serving/build_client/python/dist/ /data/mart/build_client

以上 *** 作同时解决了，如下安装过程报错

CMakeFiles/extern_snappy.dir/build.make:90: recipe for target 'third_party/snappy/src/extern_snappy-stamp/extern_snappy-download' failed
make[2]: *** [third_party/snappy/src/extern_snappy-stamp/extern_snappy-download] Error 1
CMakeFiles/Makefile2:676: recipe for target 'CMakeFiles/extern_snappy.dir/all' failed
make[1]: *** [CMakeFiles/extern_snappy.dir/all] Error 2
CMakeFiles/extern_utf8proc.dir/build.make:90: recipe for target 'third_party/utf8proc/src/extern_utf8proc-stamp/extern_utf8proc-download' failed
make[2]: *** [third_party/utf8proc/src/extern_utf8proc-stamp/extern_utf8proc-download] Error 1
CMakeFiles/Makefile2:622: recipe for target 'CMakeFiles/extern_utf8proc.dir/all' failed
make[1]: *** [CMakeFiles/extern_utf8proc.dir/all] Error 2
fatal: unable to access 'https://github.com/gflags/gflags.git/': gnutls_handshake() failed: The TLS connection was non-properly terminated.
-- Had to git clone more than once:
          3 times.
CMake Error at /home/Serving/build_server/third_party/gflags/tmp/extern_gflags-gitclone.cmake:31 (message):
  Failed to clone repository: 'https://github.com/gflags/gflags.git'


CMakeFiles/extern_gflags.dir/build.make:90: recipe for target 'third_party/gflags/src/extern_gflags-stamp/extern_gflags-download' failed
make[2]: *** [third_party/gflags/src/extern_gflags-stamp/extern_gflags-download] Error 1
CMakeFiles/Makefile2:757: recipe for target 'CMakeFiles/extern_gflags.dir/all' failed
make[1]: *** [CMakeFiles/extern_gflags.dir/all] Error 2
fatal: unable to access 'https://github.com/pybind/pybind11.git/': Failed to connect to github.com port 443: Connection timed out
fatal: unable to access 'https://github.com/madler/zlib.git/': Failed to connect to github.com port 443: Connection timed out
-- Had to git clone more than once:
          3 times.
CMake Error at /home/Serving/build_server/third_party/pybind/tmp/extern_pybind-gitclone.cmake:31 (message):
  Failed to clone repository: 'https://github.com/pybind/pybind11.git'


-- Had to git clone more than once:
          3 times.
CMake Error at /home/Serving/build_server/third_party/zlib/tmp/extern_zlib-gitclone.cmake:31 (message):
  Failed to clone repository: 'https://github.com/madler/zlib.git'


CMakeFiles/extern_pybind.dir/build.make:91: recipe for target 'third_party/pybind/src/extern_pybind-stamp/extern_pybind-download' failed
make[2]: *** [third_party/pybind/src/extern_pybind-stamp/extern_pybind-download] Error 1
CMakeFiles/Makefile2:541: recipe for target 'CMakeFiles/extern_pybind.dir/all' failed
make[1]: *** [CMakeFiles/extern_pybind.dir/all] Error 2
CMakeFiles/extern_zlib.dir/build.make:90: recipe for target 'third_party/zlib/src/extern_zlib-stamp/extern_zlib-download' failed
make[2]: *** [third_party/zlib/src/extern_zlib-stamp/extern_zlib-download] Error 1
CMakeFiles/Makefile2:730: recipe for target 'CMakeFiles/extern_zlib.dir/all' failed
make[1]: *** [CMakeFiles/extern_zlib.dir/all] Error 2
Makefile:129: recipe for target 'all' failed
make: *** [all] Error 2

还是报错，直接安装编译好的whl

whl包：
https://paddle-qa.bj.bcebos.com/PaddleServing/grpcio1.37.1/paddle_serving_app-0.8.3-py3-none-any.whl
https://paddle-qa.bj.bcebos.com/PaddleServing/grpcio1.37.1/paddle_serving_client-0.8.3-cp37-none-any.whl
https://paddle-qa.bj.bcebos.com/PaddleServing/grpcio1.37.1/paddle_serving_server_gpu-0.8.3.post101-py3-none-any.whl

bin：
https://paddle-qa.bj.bcebos.com/PaddleServing/grpcio1.37.1/serving-gpu-101-0.8.3.tar.gz

要把 Serving的bin文件拷贝到pip install paddle_serving_server的目录下，或者 export SERVING_BIN=serving bin的绝对路径

注意到上一小节的最后一行export SERVING_BIN，运行python端Server时，会检查SERVING_BIN环境变量，如果想使用自己编译的二进制文件，请将设置该环境变量为对应二进制文件的路径，通常是export SERVING_BIN= B U I L D D I R / c o r e / g e n e r a l − s e r v e r / s e r v i n g 。 其 中 B U I L D D I R 为 b u i l d s e r v e r 的 绝 对 路 径 。 可 以 c d b u i l d s e r v e r 路 径 下 ， 执 行 e x p o r t S E R V I N G B I N = {BUILD_DIR}/core/general-server/serving。 其中BUILD_DIR为build_server的绝对路径。 可以cd build_server路径下，执行export SERVING_BIN= BUILDD​IR/core/general−server/serving。其中BUILDD​IR为builds​erver的绝对路径。可以cdbuilds​erver路径下，执行exportSERVINGB​IN={PWD}/core/general-server/serving