Skip to content

v2.59.0 Post-Dencun Fixes

Compare
Choose a tag to compare
@yperbasis yperbasis released this 20 Mar 14:03
· 8 commits to release/2.59 since this release
0f0a445

Attention Caplin Users

  1. Upgrade your erigon node, let it run for about a minute, then stop it.
  2. Reset the headers & snapshots stages:
rm -rf <datadir>/caplin <datadir>/snapshots <datadir>/downloader
make integration
./build/bin/integration stage_headers --reset --datadir=<datadir>
./build/bin/integration stage_snapshots --reset --datadir=<datadir> 
  1. Start your erigon node.

More Aggressive Log/Receipt Pruning

PR #9123 introduces more aggressive pruning of logs & receipts (for nodes with r in their pruning flag, e.g. --prune=hrtc), resulting in a 2x smaller disk footprint.
This feature is only enabled for new nodes. To enable it for an existing node, you have to re-sync it from scratch.

Optional reduce IO for getTxnByHash operation

Transaction lookup by hash did touch too much places in snapshots. It caused high IO and PageCache evictions.
If you have much RPC requests - it will greatly help for this use-case.
To get it: rm datdir/snapshots/*.idx and restart erigon.
By #9506

Release Highlights

New Contributors

Full Changelog: v2.58.2...v2.59.0