Disk Growth Guidance

This document provides disk growth estimates and sizing recommendations for testnet and mainnet deployments.

Storage Components

The chain stores data in several components:

Component	Location	Growth Rate	Description
Blocks	`data/blocks/`	~2-5 KB/block	Block headers and transaction data
State	`data/state/`	Variable	Account state, contract storage
Marshal Archive	`data/marshal/`	~1-3 KB/finalization	Finality proofs and attestations
Checkpoints	`data/checkpoints/`	~10-50 MB/checkpoint	Periodic state snapshots

Estimating Growth

Per-Block Growth

block_size = header_size + (tx_count * avg_tx_size)
           = ~200 bytes + (N * ~250 bytes)

For a network with:

1 block/second
Average 10 txs/block
~2.7 KB/block

Monthly growth: ~7 GB/month raw blocks

Per-Finalization Growth

Finality proofs include BLS signatures and attestations:

~1-3 KB per finalization
With 1 finalization/second: ~2.5-7.5 GB/month

State Growth

State growth depends heavily on contract activity:

Account creation: ~100 bytes/account
Contract deployment: ~1-100 KB/contract
Storage writes: 32 bytes key + value size

Recommended Sizing

Testnet Node

Resource	Minimum	Recommended
Disk	100 GB SSD	250 GB NVMe
Expected 6-month growth	~50 GB	~50 GB
Headroom	50 GB	200 GB

Mainnet Validator

Resource	Minimum	Recommended
Disk	500 GB NVMe	1 TB NVMe
Expected 1-year growth	~200 GB	~200 GB
Headroom	300 GB	800 GB

Archive Node

Archive nodes retain all historical state:

Resource	Minimum	Recommended
Disk	2 TB NVMe	4 TB NVMe
Expected 1-year growth	~500 GB - 1 TB	~500 GB - 1 TB

Monitoring Disk Usage

Key Metrics

Monitor these Prometheus metrics to track disk usage:

# Block height (track growth rate)
rate(ashen_block_height[1h]) * 3600

# Estimated blocks per day
rate(ashen_block_height[1d]) * 86400

# Transaction throughput
rate(ashen_block_transaction_count[5m])

Grafana Dashboard

The included dashboard (ops/grafana/ashen-dashboard.json) provides:

Block height and finalized height
Transaction throughput
RPC request rates
Block production timing
Txpool size

Disk Usage Alerts

Recommended Prometheus alert rules:

groups:
  - name: disk-alerts
    rules:
      - alert: DiskSpaceLow
        expr: node_filesystem_avail_bytes{mountpoint="/data"} / node_filesystem_size_bytes{mountpoint="/data"} < 0.15
        for: 5m
        labels:
          severity: warning
        annotations:
          summary: "Disk space below 15%"

      - alert: DiskSpaceCritical
        expr: node_filesystem_avail_bytes{mountpoint="/data"} / node_filesystem_size_bytes{mountpoint="/data"} < 0.05
        for: 1m
        labels:
          severity: critical
        annotations:
          summary: "Disk space below 5%"

Pruning Strategies

Block Pruning (Non-Archive Nodes)

Keep only recent blocks (e.g., last 10,000):

# Pruning is configured via node config
--pruning-keep-recent 10000

State Snapshots

State checkpoints are stored and accessible via RPC (NodeRpcV1.list_checkpoints, NodeRpcV1.get_checkpoint).

Note: Automatic checkpoint interval configuration (--checkpoint-interval) is not yet implemented. Checkpoints are currently created during state sync and snapshot operations.

Log Rotation

Configure log rotation for node logs:

/var/log/ashen-node/*.log {
    daily
    rotate 7
    compress
    delaycompress
    missingok
    notifempty
}

Capacity Planning

Formula

required_disk = current_usage
              + (daily_growth * retention_days)
              + (headroom_factor * 1.2)

Example Calculation

For a testnet validator running 6 months:

Current usage: 20 GB
Daily growth: ~300 MB
Retention: 180 days
Headroom factor: 1.2

required_disk = 20 GB + (0.3 GB * 180) + (54 GB * 0.2)
              = 20 GB + 54 GB + 10.8 GB
              = ~85 GB minimum
              = ~125 GB recommended (with buffer)

Performance Considerations

SSD vs HDD

Required: NVMe SSD for validators
Acceptable: SATA SSD for archive/RPC nodes
Not recommended: HDD (too slow for consensus)

IOPS Requirements

Node Type	Minimum IOPS	Recommended IOPS
Validator	5,000	10,000+
RPC Node	3,000	5,000+
Archive	1,000	3,000+

Filesystem

Recommended: ext4 or XFS
Mount options: noatime,nodiratime