[mboyd1]

What Changed

• cmd/merkle-service/main.go now calls store.StartAgeSweeperFromConfig (and defers its stop), mirroring cmd/block-processor/main.go.

Why It Was Necessary

• Fixes All-in-one binary never starts the blob-store age sweeper; flat subtree store fills ext4 htree (ENOSPC at 19% disk) #267. The all-in-one process hosts the block-processor but never started the sweeper from fix(store): age-sweep orphaned subtree blobs — bounded store regardless of DAH state #190, so file:// subtree blobs that never get a delete-at-height accumulated forever. On a mainnet deployment the store reached 5.55M files in one directory, filled ext4's htree index (no large_dir), and every blob write failed with ENOSPC at 19% disk usage. That stalled SEEN_ON_NETWORK for weeks.

Testing Performed

• CGO_ENABLED=0 go build ./cmd/merkle-service, go vet ./cmd/merkle-service/, go test ./internal/store/ -run Sweep.
• The sweeper is a no-op for memory stores or when sweepIntervalSec/sweepMaxAgeSec is 0, so test and ephemeral setups are unaffected.

Impact / Risk

• Low. One sweeper per process, which is the same single-replica assumption block-processor already relies on, since all-in-one is single-process. Existing all-in-one deployments with a large backlog will see a long first sweep at startup (synchronous sweep() before the ticker). On the 5.5M-file store that would be a one-time multi-minute directory walk.

Notifications

[github-actions]

👋 Thanks, @mboyd1!

This pull request comes from a fork. For security, our CI runs in a restricted mode.
A maintainer will triage this shortly and run any additional checks as needed.

• 🏷️ Labeled: fork-pr, requires-manual-review
• 👀 We'll review and follow up here if anything else is needed.

Thanks for contributing to bsv-blockchain/merkle-service! 🚀

[mrz1836]

LGTM