Make CachedEnvironment robust to concurrent modifications
#5503
Comments
|
I think I know what to do here but it's slightly tricky. We might be able to reuse it for tool installs though. We want to create the environment in one location, then relocate it to the final location (probably by symlinking so that it's atomic). But we can't do this naively because we need the interpreter in any installed scripts to use the path of the final location. |
charliermarsh
added a commit
that referenced
this issue
Jul 29, 2024
…nts (#5509) ## Summary The idea here is similar to what we do for wheels: we create the `CachedEnvironment` in the `archive-v0` bucket, then symlink it to its content-addressed location. This ensures that we can always recreate these environments without concern for whether anyone else is accessing them. Part of the challenge here is that we want the virtual environments to be relocatable, because we're now building them in one location but persisting them in another. This requires that we write relative (rather than absolute) paths to scripts and entrypoints. The main risk with relocatable virtual environments is that the scripts and entrypoints _themselves_ are not relocatable, because they use a relative shebang. But that's fine for cached environments, which are never intended to leave the cache. Closes #5503.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Right now, it's possible that we'd modify an environment that's in-use by another process.
The text was updated successfully, but these errors were encountered: