feat: dockerfile optimization cases

jina-ai / executor-normalizer

Jina executor package normalizer

5 stars 0 forks source link

feat: dockerfile optimization cases #13

Closed numb3r3 closed 3 years ago

numb3r3 commented 3 years ago

Collection of optimization cases

Optimization

[ ] always use jina as base image https://jinaai.slack.com/archives/C018QC78L84/p1625065827203300
[x] remove unecessary lines https://jinaai.slack.com/archives/C023A7RAPGR/p1625126855404700

Bugs/Features

[x] use py-version specific base image jinaai/jina:2.0.3-py37
[x] check if the files are just empty placeholder https://jinaai.slack.com/archives/C023A7RAPGR/p1625199017424900
[x] complete requirements.txt https://github.com/jina-ai/executor-normalizer/pull/18

numb3r3 commented 3 years ago

always use jina as base image https://jinaai.slack.com/archives/C018QC78L84/p1625065827203300

To approach it, we need first answer the following questions: 1) how to handle the case of using cuda drivers ? 2) how to make the Dockerfile optimization safe?

some optimization/revisions on Dockerfile would cause some errors. It is very hard to make sure the optimized Dockerfile can work perfectly and produces less image size

CatStark commented 3 years ago

We wont continue with " always use jina as base image"