Opera blockchain protocol secured by the Lachesis consensus algorithm
Go-opera 1.1.3-rc.5 is a small optimization release aimed to fix a performance issue.
Go-opera 1.1.3-rc.4 is an optimization release to enhance performance.
This update includes an enhanced event emission algorithm, which reduces Time to Finality (TTF) without increasing the number of emitted events per second.
For API nodes, the update is recommended. It addresses previously identified Out-Of-Memory (OOM) crashes in the filter API and enhances the estimation of gas tip
/gas price
for transactions.
gas tip
/gas price
estimation (#490, #492)Full Changelog: https://github.com/Fantom-foundation/go-opera/compare/v1.1.2-rc.6...v1.1.3-rc.4
Pre-release for Go-opera 1.1.3
Go-opera 1.1.2-rc.6 is a maintenance release. It includes the following fixes:
db heal
when combined with fully pruned EVM (e.g. with --gcmode=full
) (#432)--db.migration.mode rebuild
migration and for db transform
command (#435)1.1.2-rc.5 is an optimization release. Improvements include batched genesis blocks processing, configurable DBs management, parallel EVM logs search and optimizations in p2p protocol. Depending on hardware, it can achieve approximately ~1.3x improvement for events/blocks processing, ~3x improvement for genesis file processing (except for archive genesis files), ~5x improvement for logs search.
There are two migration options for upgrading to 1.1.2-rc.5:
--db.migration.mode reformat --db.preset legacy-ldb
. This option is instant but will leave out some of the performance improvements for blocks/events processing.--db.migration.mode rebuild --db.preset X
, where X is a selected DBs layout. Migration will take a lot of time - for an NVMe SSD drive, the number of hours would be roughly datadir size in GB
divided by 50 GB/hour
. If the remaining disk space is insufficient to store a copy of largest DB in datadir/chaindata, then the migration will take up to 3 times longer, depending on available disk space.Flags --db.migration.mode
is required only if datadir isn't migrated yet.
There are 2 options for the new --db.preset
flag:
ldb-1
- fastest LevelDB layout (default option).pbl-1
- fastest Pebble layout.Pebble offers slightly better performance and smaller IO utilization on average, especially with extended cache. We don't recommend using Pebble presets for validators nor other mission-critical nodes yet.
--rpc.evmtimeout
flag (#393)v1.1.1-rc.2 is a hotfix release for an issue. The upgrade is highly recommended for all nodes.
This release features automatic EVM pruning which can be enabled with either --gcmode full
or --gcmode light
flag. --gcmode full
option will prune much more EVM nodes than --gcmode light
at expense of worse performance.
0
into API output for web3 compatibility (#333)Release v1.1.1-rc.1 allows new read nodes to use new customizable genesis files and to sync using snapsync
mode.
The upgrade is highly recommended for API nodes, as it fixes race condition in filter API calls and enables improved gas price (priority fee) estimation. The upgrade is not recommended for validator nodes yet.
Note that, if you're running a 1.1.0-rc.X, after the upgrade to 1.1.1-rc.1, it won't be possible to use the old genesis files (mainnet.g
and testnet.g
). Instead, you can omit --genesis
flag and you don't need to download a new genesis file (as the file is no longer mandatory for already initialized nodes).
Release notes:
sfc
API namespace
-exitwhensynced
flag (#304)v1.1.0-rc.5 is a hotfix release for API nodes. It adds support of eth_feeHistory
API method required by some EIP-1559 clients.
Release notes:
The releases 1.1.x enable the support of London and LLR hard-fork upgrades.
Release notes: