chore(server/cuda): bump pinned llama.cpp digest to 2ea4be99 - #89
Merged
Conversation
Upstream ghcr.io/ggml-org/llama.cpp:server-cuda moved past the pinned digest. Bump 67bf4625 → 2ea4be99 (current latest, verified via imagetools inspect). The digest is pinned because the image layout has shifted before (the libllama-server-impl.so split that took prod down with exit 127) — so this must be validated with a scout-cuda rebuild + a GPU smoke test that llama-server resolves all libs and embeddings actually run on the GPU before it goes anywhere near a release. Addresses the weekly llama.cpp-pin-freshness issues #84 #83 #82 #58. Co-Authored-By: Claude Opus 4.8 <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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Bumps the pinned
ghcr.io/ggml-org/llama.cpp:server-cudadigest inserver/Dockerfile.cuda— upstream moved past the previously-pinned image.sha256:67bf4625…94sha256:2ea4be99…41(current upstreamserver-cuda, verified viaimagetools inspect)The digest is pinned deliberately because the image layout has shifted before (the
libllama-server-impl.sosplit that took prod down withllama-serverexit 127). So this was validated end-to-end on the RTX 3090 before opening the PR.Validation (dev image
:cu128-dev, built from this branch)make scout-cuda→ 0C / 0H / 0M / 0LCOPY --from=llama-source /app/*.so* /app/succeeded → new layout still provides the libscix) and verified:RestartCount 0, not OOMKilled/health → 200/app/llama-serverlaunched and stayed up with--n-gpu-layers 99 --cache-ram 0(exit-127 lib-break did not recur)498 MiB), not CPU fallbackIssues
Addresses the weekly
llama.cpp-pin-freshnessissues #84, #83, #82, #58 (duplicates of the same bump). They'll close when this reachesmainwith the next release; will close manually otherwise.Note
The dev box is currently running this build (
:cu128-dev). On the next release it gets promoted to:cu128/:vX.Y.Z-cu128via the normal tag flow; rollback meanwhile is redeploying:v0.11.0-cu128.🤖 Generated with Claude Code