-
Notifications
You must be signed in to change notification settings - Fork 471
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Disable docker sandbox #4649
Merged
Merged
Disable docker sandbox #4649
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This reverts commit 661ab6d.
stgpetrovic
force-pushed
the
disable-docker-sandbox
branch
4 times, most recently
from
February 17, 2023 13:57
2db515b
to
fa0cd25
Compare
stgpetrovic
force-pushed
the
disable-docker-sandbox
branch
from
February 17, 2023 14:35
fa0cd25
to
db0026a
Compare
will-cromar
approved these changes
Feb 17, 2023
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for fixing this!
yeounoh
approved these changes
Feb 17, 2023
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thank you!
mateuszlewko
pushed a commit
that referenced
this pull request
Mar 15, 2023
* Revert "Disable the Sandbox build in Docker (#4515)" This reverts commit 661ab6d. * Remove submodule step from TPU CI * Update cloudbuild.yaml * Remove separate TF build step * Add bazel WORKSPACE and rc * Update Dockerfile * throw more compute at the TPU CI build * more bazel jobs * Revert to old machine type and default to local builds. * Revert to default nonsandbox * Remove jobs limitation * Try smaller shm * Fix build of llvm in sandbox * Mount 7g of ram properly * TEMP remove pytorch build * remove install of packages * confirm shm creation * Revert machine type --------- Co-authored-by: Will Cromar <[email protected]> Co-authored-by: Will Cromar <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Revert to local builds on cloud vs sandboxed (shm size issues).
There are a couple of issues there that intermingle.
Bazel has a bug where the cppmap files don't get included as deps when building with clang when not in sandbox, from an external repository (local repository is OK). Using a cache might alleviate this.
The suggested solution is building in the sandbox.
For some reason,
docker build --shm-size=16g
does not prduce a 16g/dev/shm
(it is actually 64M which is the default). There are differences in docker build/run but during build it should have 16g but it does not (as checked withRUN df -h
showing 64M).So the solution for this issue is to build sandboxed (without cache, unlike the other build that has cache and does not suffer from this; we could replicate that in tpu tests too), and in docker, but to run with
RUN --mount=type=tmpfs,target=/dev/shm,rw cmd
which then mountstmfps
as/dev/shm
and builds properly. The build needs about4G
of temp files so16G
is an overkill there.This makes the build work (avoid bazel bug by sandboxing), and makes it fast (avoid sandbox base being on HDD, and avoid docker
shm-size
issue by mounting thetmpfs
).We can rework this as needed in the future when we streamline all different builds into a single config.