Skip to content

Operations Small Host

github-actions[bot] edited this page Nov 21, 2026 · 6 revisions

Running Hyphen on a small host

The earlier Linux deployment had 4 vCPUs, 2 GiB of RAM and a 40 GB disk. Its measurements predate HYN2. In a Windows HYN2 smoke run, the peak working sets before restart were 38,092,800 bytes for the node, 10,784,768 for the pool and 7,233,536 for the miner. The sample covered a fresh chain at height 3. The wallet scan peaked at 5,095,424 bytes. Longer histories, larger ledgers and more peers can increase memory use. Reproduction

Verification cost

A full HYN3 block verification derives the weights, evaluates one nonce and performs up to one 1 MiB BLAKE3 scratchpad calculation. Measure the node, miner and pool separately at the chosen concurrency. Workload

Storage cache

sled defaults to a one gigabyte page cache. The node sets it explicitly:

hyphen-node ... --db-cache-mb 64     # default; 16 is the floor

Mining on the same host

Do not. Two mining threads produced ~256 seconds of CPU time in a 130-second window; on four cores that starves the node, the pool and the tunnel, which is how a healthy origin starts missing tunnel keepalives. --threads 0 leaves one core free, but on four cores that is still 75% of the machine.

Mine elsewhere and connect over TCP.

Retention and storage budgets

Structure Bound
Mempool entries 10 000 entries, byte budget derived from that
Explorer miner history 512 blocks per reward key, 4096 keys
Explorer index batches 256 blocks per batch, 8 batches per request
Pool payout history Target of 512 records. Reserved and submitted payouts stay until reconciled.
Pool reward events / recent blocks 512 / 128

Block size and fee priority

max_block_size is part of consensus_params_hash, so changing it produces a new chain identity: a fresh data directory, a rebuilt pool and miner, and a genesis hash every participant must agree on. Raise it deliberately, never as a silent edit.

Within a block, space is allocated in two passes: protocol-reserved entries take up to a quarter of the block, then ordinary transactions are packed in fee-density order. A transaction that pays more per byte is included before one that pays less, which is what makes a contract deployment compete for space the same way a transfer does.

When no block appears

The pool answers "what is the mining loop doing right now" in one read:

curl -fsS http://127.0.0.1:8081/api/pool/status

Use the status response and current tip together to diagnose stalled production.

What it shows What it means
connected_miners: 0 Nothing authenticated. Check the miner's --pool host/port and that the pool is listening on the interface the miner reaches — not the node, not the tunnel.
job is absent No template has arrived from the node yet. The pool does nothing at all in this state.
job.age_secs pinned well above the 30 s block target, blocks.accepted flat The pool is still handing out a job the node has already replaced. Look at the node and the path to it (--node), not at the miners.
connected_miners > 0, shares.accepted flat HYN3 SOLO counts full block candidates. Check each miner's current job, scoring rate, target and rejection log.
last_rejection: "stale job" A candidate arrived for a job another miner had already turned into a block. Normal in bursts, especially on a devnet ramp where the first blocks are near-instant; a constant stream means miners are racing on a job that should have been replaced.
blocks.submitted > 0, blocks.accepted: 0 The node refused every block. last_block.detail carries the node's own reason — read that before touching anything else.
template_refresh_failures climbing The node did not serve a template after a block was accepted. The pool retries four times and then waits for the node push; a persistent count is a node problem.

shares.accepted counts candidate submissions this pool accepted. Scan the canonical chain with the wallet to verify a SOLO reward and its confirmations.

Recommended layout

  • Node, pool and tunnel on the 4 vCPU / 2 GiB box.
  • Miner elsewhere.
  • --db-cache-mb 64 and the unit's MemoryMax=1600M (already in deploy/systemd/hyphen-node.service; override with systemctl edit hyphen-node if your host needs a different ceiling) so the kernel reclaims from the node rather than from the tunnel.
  • Cloudflare tunnel as its own service, see deploy/systemd/cloudflared.service.example.
  • Alert on /healthz (liveness) and on df for the data directory.

Clone this wiki locally