docs: refresh disk sizes on main, add sepolia, hoodi and chiado - #22867
Open
awskii wants to merge 4 commits into
Open
docs: refresh disk sizes on main, add sepolia, hoodi and chiado#22867awskii wants to merge 4 commits into
awskii wants to merge 4 commits into
Conversation
main's copy is 11 months stale (2025-09-01): the mainnet full-node figure reads 920 GB against a measured 419.04 GB. update-disk-sizes.yml pins its base to release/3.5 and never writes main, so main only ever catches up by hand. Values copied from release/3.6 (measured 2026-07-19 mainnet, 2026-07-21 gnosis); hardware-requirements.mdx and the llms artifacts are regenerated from the JSON.
Only mainnet and gnosis were listed. Archive sizes measured on the snapshotter-bm-v36-* nodes on 2026-07-29: sepolia 1.10 TB, hoodi 133.73 GB, chiado 29.64 GB. Execution-layer data only, so the tables carry an archive row and say so; full and minimal are not measured for these networks.
awskii
requested review from
AskAlexSharov,
bloxster and
yperbasis
as code owners
July 29, 2026 14:11
The breakdown was a Nov 2024 mainnet-only snapshot, and it left ~1.2 TB unaccounted for by omitting the block segments that sit in snapshots/ itself. Replaced with a table covering eth-mainnet, gnosis, sepolia, hoodi and chiado, measured on the snapshotter-bm-v36-* archive nodes in July 2026, with block segments broken out and the off-by-default data named.
architecture.md said chaindata is ~15 GB on mainnet and the landing stat claimed an archive node is under 2 TB. Measured on a 3.6 mainnet archive node: 22.75 GB chaindata, 2.17 TB total.
awskii
enabled auto-merge
July 29, 2026 15:13
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
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.
docs/site/src/data/disk-sizes.jsonon main is 11 months stale (2025-09-01): it reads 920 GB for a mainnet full node against a measured 419.04 GB.update-disk-sizes.ymlpinsbase_branchtorelease/3.5, so main is never written and only ever catches up by hand. The published site builds from the release branch, so it already serves the current values — this brings main in line.Sepolia, Hoodi and Chiado had no entry anywhere, and the datadir breakdown on the Database page was a Nov 2024 mainnet-only snapshot that left ~1.2 TB unaccounted for by omitting the block segments in
snapshots/itself.Changes
disk-sizes.jsonand the hardware-requirements tabsarchitecture.md(chaindata~15 GB -> ~20 GB) and the landing stat (< 2 TB->~2 TBfor an archive node)hardware-requirements.mdxand thellms-full.txtartifacts from the JSON viarender-disk-sizes.py/generate-llms.pyNew figures are measured on the
snapshotter-bm-v36-*archive nodes on 2026-07-29. Those boxes run with commitment history, the receipts cache and the Caplin archive enabled, so the optional components are excluded per file rather than by subtracting a lump —erigon snapshots dureproduces the same per-category totals, and gnosis, the one node without commitment history, agrees to 0.05%.Notes
The three new networks carry an archive row only, and their captions say so: the CI arms that measure full and minimal (
qa-sync-from-scratch-full-node.yml,qa-sync-from-scratch-minimal-node.yml) both pinchain: [mainnet, gnosis]. Adding sepolia/hoodi there is the only way to get real numbers, and it is a QA-capacity call rather than a docs one.update-disk-sizes.ymlis unchanged, so main will drift again. Retargeting it is a separate call about which branch is the docs source of record.versioned_docs/version-v3.3is also unchanged, which is correct for a frozen snapshot.