serve: serialize credential_process with per-profile flock#31
Merged
jgowdy-godaddy merged 1 commit intomainfrom Apr 17, 2026
Merged
serve: serialize credential_process with per-profile flock#31jgowdy-godaddy merged 1 commit intomainfrom
jgowdy-godaddy merged 1 commit intomainfrom
Conversation
Two AWS CLI invocations firing `credential_process = awsenc serve` in parallel on the same profile, with the cache in Refresh or Expired state, would each independently run the transparent-reauth chain (Okta-session decrypt → SAML → STS AssumeRoleWithSAML) or fall back to printing stale cached creds. Both writers then raced the cache update. Duplicate STS traffic, possible Okta rate-limit hit, and the losing writer's fresh creds silently overwritten. Take an exclusive advisory lock on `<cache>.lock` before reading the cache, hold it across all state branches (Fresh / Refresh / Expired), and release on drop. A second caller blocks at lock_exclusive, waits for the first to finish, then re-reads the now-refreshed cache and prints its credentials — single STS call, single prompt chain, consistent session_start. Uses fs4 for cross-platform flock / LockFileEx. The lock file is a zero-byte sidecar that stays on disk; crash recovery is the OS's job (inode lock released on handle close).
jgowdy-godaddy
pushed a commit
that referenced
this pull request
Apr 17, 2026
main's #31 landed a per-profile serve flock that overlaps my own flock commit. Their implementation is nearly identical; I take theirs (serve_lock_path_for_cache + matching ServeLock) and drop the duplicate serve_lock_path + ServeLock definitions from my branch. Call sites use theirs' slightly more tolerant cache::cache_path().map().unwrap_or_else() fallback. Our remaining unique changes on top of main: - main.rs dispatch: validate profile arg before create_storage (fail-fast on serve/exec without --profile, cleaner error message, avoids popping Keychain prompts on malformed invocations) - cache.rs: APL1 envelope wrap_for_encrypt / unwrap_after_decrypt + counter sidecar helpers - auth.rs: header-first then wrap-each-secret ordering, counter persistence after write - serve.rs: decrypt_aws_credentials_with_envelope used by the Fresh + Refresh-fallback paths; transparent-reauth re-wraps under the new header - exec.rs: unwrap_after_decrypt in get_cached_credentials
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
2 participants
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.
Summary
Two AWS CLI invocations firing `credential_process = awsenc serve` in parallel on the same profile, with the cache in Refresh or Expired state, would each independently run the transparent-reauth chain (Okta-session decrypt → SAML → STS AssumeRoleWithSAML) or fall back to printing stale cached creds. Both writers then raced the cache update. Result: duplicate STS traffic, possible Okta rate-limit hit, and the losing writer's fresh credentials silently overwritten.
Take an exclusive advisory lock on `.lock` before reading the cache, hold it across all state branches (Fresh / Refresh / Expired), and release on drop. A second caller blocks at `lock_exclusive`, waits for the first to finish, then re-reads the now-refreshed cache and prints its credentials — single STS call, single prompt chain, consistent session_start.
Uses `fs4` for cross-platform `flock` / `LockFileEx`. The lock file is a zero-byte sidecar that stays on disk; crash recovery is the OS's job (inode lock released on handle close).
Test plan