Limit the memory of a dockerized test job - #317
Merged
Conversation
test.py caps the address space of one test at a time, so nothing bounds the sum of the tests it runs in parallel. That is worst for wasm-jit, whose tests are allowed 16 GB of address space each because the JIT reserves ~4 GB of it per wasm memory. Cap the container the job runs in with `docker run --memory=N --memory-swap=N`, N = 85% of the node's RAM, since these jobs are pinned to the high-memory nodes. A cgroup limit charges memory in use rather than address space, covers every process test.py spawns, and on a breach the kernel kills the greediest omc - so a model can now fail because of a model tested in parallel with it, which the log makes visible: the limit is printed before the run and the peak plus the OOM kill count after it. Only the targets that build their own image (wasm-jit today) get this; the jobs running directly on a node are unchanged. Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
test.py caps the address space of one test at a time, so nothing bounds the sum of the tests it runs in parallel. That is worst for wasm-jit, whose tests are allowed 16 GB of address space each because the JIT reserves ~4 GB of it per wasm memory.
Cap the container the job runs in with
docker run --memory=N --memory-swap=N, N = 85% of the node's RAM, since these jobs are pinned to the high-memory nodes. A cgroup limit charges memory in use rather than address space, covers every process test.py spawns, and on a breach the kernel kills the greediest omc - so a model can now fail because of a model tested in parallel with it, which the log makes visible: the limit is printed before the run and the peak plus the OOM kill count after it.Only the targets that build their own image (wasm-jit today) get this; the jobs running directly on a node are unchanged.