ownJune 28, 2021
docker容器下配置jupyter notebook: https://blog.csdn.net/leng_yan/article/details/87208363 利用Docker环境配置jupyter notebook服务器: https://blog.csdn.net/eswai/article/details/79437428 解决远程问题： https://blog.csdn.net/Accepted_Lam/article/details/103837677
GitHub fchollet / deep-learnig-with-python-notebooks
Jetson AGX Xavier
git repository
samuelwei/DeepLearning/PythonDeepLearning/jupyternotebook/build/
$ sudo docker build -t pdl_jupyternotebook:r32.5.0-tf2.3-py3 -f DeepLearning/PythonDeepLearning/build/Dockerfile .
#在哪个文件夹内执行此命令，那么context 会推送本目录下的文件作为context文件到Docker remote端
$ sudo docker build -t pdl_jupyternotebook:r32.5.0-tf2.3-py3 .
-t <ImageName, pythondeeplearningjupyternotebook:r32.5.0-tf2.3-py3>
Use format:
docker build [选项] <上下文路径/URL/->
July 10, 2021 on Jetson NX
$ NV_GPU=0 nvidia-docker run -it --name pdl_jupyternotebook -p 2888:8888 -v /home/workspace/DeepLearning/:/home/workspace/DeepLearning/ -w /home/workspace/DeepLearning/PythonDeepLearning/ --restart=always samuelwei/pdl_jupyternotebook:r32.5.0-tf2.3-py3 /bin/bash
NV_GPU=0 nvidia-docker NV_GPU=0 选择第0个GPU设备 用nvidia-docker才能启用GPU 如果不用GPU，直接用docker命令就行
-it -i 以交互模式运行容器，通常与 -t 同时使用 -t 为容器重新分配一个伪输入终端，通常与 -i 同时使用
启动设备就启动容器 --restart=always
–name 后面跟容器的名称，notebook-server是我自定义的名字，可以随便改；
-p 端口映射，如xxxx:yyyy将主机的xxxx端口映射到容器内的yyyy端口，jupyter notebook默认使用8888端口，7777是我自定义的，不冲突就好；
-v 路径映射 /a : /b将主机的/a路径映射到容器的/b，根据自己的需要手动修改 --volumes-from dbdata 使用数据卷容器dbdata，另外一个容器专门存放数据使用
eswai/tf140py2:1.0是我使用的镜像名称和版本号，根据具体需要修改
启动镜像后执行的命令，即启动容器内的命令 /bin/bash /jupyter notebook &
Always restart:
$ sudo docker update --restart=always <containerID>
May 14, 2022 on Jetson Xavier
$ sudo NV_GPU=0 nvidia-docker run -it --name dl_jupyternotebook -p 2888:8888 -v /home/workspace/DeepLearning/:/home/workspace/DeepLearning/ -w /home/workspace/DeepLearning/ --restart=always samuelwei/pdl_jupyternotebook:r32.5.0-tf2.3-py3 /bin/bash
July 12, 2022 on Jetson old_AGX
$ sudo NV_GPU=0 nvidia-docker run -it --name pdl_jupyternotebook -p 2288:22 --privileged=true -v /home/workspace/DeepLearning/:/home/workspace/DeepLearning/ -w /home/workspace/DeepLearning/ --restart=always samuelwei/pdl_jupyternotebook:r32.5.0-tf2.3-py3 /bin/bash
https://blog.csdn.net/eswai/article/details/79437428
root@5805ad32505e:/home/workspace/DeepLearning#
$ cd /dbdata/DeepLearning/PythonDeepLearning/ #格式 $ cd <映射目标>
$ mkdir jupyternotebook #格式 $ mkdir <存放notebook的文件夹>
此时相当于在主机的/home/eswai下创建了jupyter文件夹
$ pip3 install jupyter notebook
$ ipython
output:
In [1]: from notebook.auth import passwd
In [2]: passwd()
Enter password:
Verify password:
Passwords do not match.
Enter password: wmp797007
Verify password: wmp797007
Out[2]: 'argon2:$argon2id$v=19$m=10240,t=10,p=8$nV3BGCYeuvxTJPWMvCHWiw$vFkidx/zXHvwMNtQQ4yeIA'
In [3]:
execute:
$ jupyter notebook --generate-config
output:
Writing default config to: /root/.jupyter/jupyter_notebook_config.py
Change the configurate:
$ vim /root/.jupyter/jupyter_notebook_config.py
修改内容如下
# 允许root启动
c.NotebookApp.allow_remote_access = True #允许远程访问
c.NotebookApp.allow_root = True
c.NotebookApp.ip='*'#×允许任何ip访问
c.NotebookApp.open_browser = False
c.NotebookApp.password = 'argon2:$argon2id$v=19$m=10240,t=10,p=8$5reTOnFAQh6t6+HrgtSFFA$Q4nJ1raguPZ1Ycj3MQdgnw'
c.NotebookApp.port =8888 #可自行指定一个端口, 访问时使用该端口
在主机终端下开启： Usage: sudo docker exec jupyter notebook &
$ sudo docker exec ce4d755e41e5 jupyter notebook &
$ sudo docker exec pdl_jupyternotebook jupyter notebook &
在容器下开启：
$ jupyter notebook --allow-root
$ jupyter notebook &
$ Ctrl+c #退出jupyter
按Ctrl+P+Q退出容器但不关闭
打开链接： http://[主机IP]:[配置映射的主机端口]/tree?token=[自定义的token]
May 14, 2022 update Xavier http://sharebee88.qicp.vip:28890/
August 20, 2021 NX
192.168.0.102:2888
http://sharebee.wicp.top:57184/
passwd: sharebee
output error:
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -DUSE__THREAD -DHAVE_SYNC_SYNCHRONIZE -I/usr/include/ffi -I/usr/include/libffi -I/usr/include/python3.6m -c c/_cffi_backend.c -o build/temp.linux-aarch64-3.6/c/_cffi_backend.o
c/_cffi_backend.c:15:10: fatal error: ffi.h: No such file or directory
#include <ffi.h>
^~~~~~~
compilation terminated.
error: command 'aarch64-linux-gnu-gcc' failed with exit status 1
solving method:
https://raspberrypi.stackexchange.com/questions/94695/failed-building-wheel-for-cffi-on-model-3b
ciffi depends on libffi, so I had to first install the libffi-dev package. Install it using:
$ sudo apt install libffi-dev
The package might be different if you are using some other distro.
add apt install libffi-dev to Dockerfile command RUN
output error:
Building wheels for collected packages: argon2-cffi, cffi
Running setup.py bdist_wheel for argon2-cffi ... error
Complete output from command /usr/bin/python3 -u -c "import setuptools, tokenize;__file__='/tmp/pip-build-cg8bc4i7/argon2-cffi/setup.py';f=getattr(tokenize, 'open', open)(__file__);code=f.read().replace('\r\n', '\n');f.close();exec(compile(code, __file__, 'exec'))" bdist_wheel -d /tmp/tmplaftfnq8pip-wheel- --python-tag cp36:
ERROR: Failed to build one or more wheels
Traceback (most recent call last):
File "/usr/local/lib/python3.6/dist-packages/setuptools/installer.py", line 75, in fetch_build_egg
subprocess.check_call(cmd)
File "/usr/lib/python3.6/subprocess.py", line 311, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['/usr/bin/python3', '-m', 'pip', '--disable-pip-version-check', 'wheel', '--no-deps', '-w', '/tmp/tmpmjdecu1n', '--quiet', 'cffi']' returned non-zero exit status 1.
solving method:
https://raspberrypi.stackexchange.com/questions/94695/failed-building-wheel-for-cffi-on-model-3b
ciffi depends on libffi, so I had to first install the libffi-dev package. Install it using:
$ sudo apt install libffi-dev
The package might be different if you are using some other distro.
add apt install libffi-dev to Dockerfile command RUN