Questions about running an archive node

[fortuna502]

Hello Geth Team,

I am setting up an archive node using Geth and would appreciate guidance on the following aspects:

1. Benchmark Tests and Performance Metrics:

• Are there any official or community-shared benchmark tests for running an archive node with Geth?
• Specifically, what is the typical Queries Per Second (QPS) for RPC methods like eth_call on an archive node? If there are no standard metrics, could you point me to resources or methods for benchmarking this myself?

2. Syncing an Archive Node:

• Does running an archive node with Geth require syncing from block 0 (genesis block)? Or can it start from a more recent block while still maintaining full historical state?
• How long does the initial sync typically take for an archive node, assuming standard hardware (e.g., 16-core CPU, 64GB RAM, and high-performance SSD)?
• Are there any recommended sync modes or flags (e.g., --syncmode full) for optimizing the sync process?

3. Archive Snapshots:

• Does Geth provide any pre-synced archive snapshots that can be used to speed up the initial setup of an archive node?
• If not, are there third-party services or community resources that offer such snapshots for Geth?

Thank you for your assistance. I look forward to your insights and recommendations.

[rjl493456442]

Does running an archive node with Geth require syncing from block 0 (genesis block)? Or can it start from a more recent block while still maintaining full historical state?

If you want to keep the full state histories since genesis, then run the full sync from genesis in gcmode=archive.

[rjl493456442]

How long does the initial sync typically take for an archive node, assuming standard hardware (e.g., 16-core CPU, 64GB RAM, and high-performance SSD)?

The legacy hash mode archive node will take very long time to sync, probably a few months. We will ship the new archive node after the fork and 2 weeks are sufficient for the sync.

[rjl493456442]

Are there any recommended sync modes or flags (e.g., --syncmode full) for optimizing the sync process?

also go with gcmode archive

[rjl493456442]

Does Geth provide any pre-synced archive snapshots that can be used to speed up the initial setup of an archive node?
If not, are there third-party services or community resources that offer such snapshots for Geth?

no

[fortuna502]

Hi @rjl493456442 , Thank you for the initial insights. I have a few follow-up questions to better understand running an archive node with Geth:

1. Benchmarks for Archive Nodes:
   Are there any specific benchmark tests available for archive nodes, either from the Geth team or the community? For example, metrics on CPU, memory, or disk I/O usage under typical workloads would be helpful.

2. JSON-RPC QPS:
   What is the expected Queries Per Second (QPS) for JSON-RPC interfaces, such as eth_call, on an archive node? If no official metrics exist, are there recommended tools or methods to measure this?

3. Sync Efficiency (Archive vs. Full Node):
   Does the sync process for an archive node differ significantly in efficiency compared to a full node? For instance, does the need to store all historical states slow down block synchronization noticeably?

4. Sync Lag in Archive Nodes:
   Is there a risk of an archive node falling behind in block synchronization (e.g., lagging behind the latest blocks)? If so, what factors contribute to this, and how can it be mitigated?

I appreciate any guidance or references to relevant resources. Thank you!