Hacker News new | past | comments | ask | show | jobs | submit login
macOS and Docker volumes – chmod and performance?
1 point by XAKEPEHOK 88 days ago | hide | past | favorite
I’ve been working on Linux for several years and now planning to switch to Mac. I’m already familiar with ARM images and x86 emulation nuances, so I understand that everything should generally work. However, I’m looking for insights into specific edge cases, especially when comparing to Windows, since Linux is straightforward in these areas.

Questions:

1. Performance on file systems: On Windows, with WSL2, there are significant performance issues at the boundary of the host filesystem and WSL.

If the project resides inside WSL’s filesystem, applications run fast, but IDEs (JetBrains) lag due to file sync overhead.

Conversely, if the project is on the host machine and mounted in WSL, IDEs perform well, but the application runs painfully slow. Are there any similar performance issues on Mac? My question is not about specific IDEs or their configurations, nor about WSL, but rather about containerization and file system behavior on macOS.

2. Volumes and permissions: On Windows, all files and directories inside Docker volumes have chmod 777 within the container. While this works, it’s far from ideal.

I prefer setting specific uid/gid and user configurations in my containers to align dev environments with production, ensuring no surprises like missing permissions or needing root access to back up volumes on the host machine.

I don’t care about the exact chmod values on Mac as long as I can, for example, delete files from volumes without sudo and have the container enforce its own file permissions and ownership. Does Mac allow for this level of control?

3. Performance with volumes: ChatGPT suggested that there might be significant performance drops when storing MongoDB or Elasticsearch data inside volumes on Mac.

I’m not setting up production on a Mac, but I plan to use it for data processing tasks, where datasets in MongoDB or Elasticsearch could reach 10 GB.

How noticeable is the slowdown in practice? I’m looking for subjective feedback rather than exact metrics.

This question is aimed at experienced users who understand the nuances of containerization. Please don’t suggest going back to Windows/WSL2 or using VSCode, etc. My knowledge of WSL2 may be outdated, and Windows might already perform better, but my focus right now is on Mac.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: