Skip to content

Add bounded shard clients and atomic membership fencing - #11

Merged
loverustfs merged 2 commits into
mainfrom
codex/bounded-shard-client
Oct 21, 2026
Merged

loverustfs merged 2 commits into
mainfrom
codex/bounded-shard-client

Conversation

@loverustfs

@loverustfs loverustfs commented Oct 20, 2026 •

Copy link
Copy Markdown
Contributor

Change

Add ShardedClient with full-identity rendezvous routing, bounded connection/retry tables, per-shard sessions and stale-reply suppression. Membership replacement validates the entire snapshot and fences all affected local state before publishing it; failed fencing leaves membership and credit unchanged. Client statistics distinguish logical descriptor bytes from wire traffic.

This branch includes the TCP reply-budget fix from #10. Route publication is local to a client and is not a remote-server fencing mechanism; operators still coordinate ownership changes.

Validation

Run the retained CPU gate with a fresh evidence directory:

python3 scripts/ci/vm_regression.py --profile cpu --offline --output artifacts/merge-review-cpu

Fetch locked dependencies once before the offline run. The runner checks formatting, default/all-feature Clippy, workspace/doc tests, optional-feature refusal contracts and CPU demos; it writes results.json, environment.json, summary.md, JUnit and raw logs. GitHub CI runs Linux x86_64, native aarch64 and aarch64 qemu-user checks and uploads x86_64 regression evidence.

Boundaries

GCK1/GCKR and CacheIdentity remain unchanged. Missing, failed or unverified blocks require recomputation. No hot erasure coding, durability replicas, linked accelerator libraries or product performance samples are added. TCP remains a trusted-fabric plaintext transport; the bearer fingerprint does not provide mTLS.

Real RDMA/NIXL, GDS, BF3/DOCA/SPDK, vLLM export/import and live RustFS remain NOT_RUN_HARDWARE. GPU/model execution remains NOT_RUN_BUDGET; an approved model/license campaign remains NOT_RUN_LICENSE. TCO_UNKNOWN, go_gate=NOT_EVALUATED, Mooncake NOT_LINKED and NOT_A_PERFORMANCE_RESULT remain.

@loverustfs
loverustfs changed the base branch from codex/hot-tcp-admission to main October 21, 2026 00:49
@loverustfs
loverustfs marked this pull request as ready for review October 21, 2026 00:49
@loverustfs
loverustfs merged commit fd4ea48 into main Oct 21, 2026
6 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant