IMPORTANT NOTE: The Trace
data are incompatible with v1.1.1
or older versions, and the old data might be incorrect. If you were running a node with executive_trace=true
, please clear the data and resync the chain.
Improvements
Configuration Improvements
- Change the default node type to
full
node instead ofarchive
node. And allow setting the node type in the
configuration file with the entrynode_type
. - Add parameters to independently configure the garbage collection time of different kinds of data (like receipts,
transactions, block traces, state, e.t.c.). Check theadditional_maintained_*
entries inrun/tethys.toml
. - If
block_db_dir
ornetconf_dir
is not set, put the default directory in the one configured withconflux_data_dir
.
The old behavior is to be put in the hard-coded./blockchain_data
. - Add a parameter
public_rpc_apis
to control the publicly available RPC interface sets.
The access totest
anddebug
RPCs is no longer related tomode
. - Remove the parameter
enable_tracing
because it has been included in the newpublic_rpc_apis
.
RPC Improvements
- Add new local RPC
cfx_getEpochReceipts
to allow querying receipts based on an epoch number. - Add new trace RPC
trace_filter
to allow querying traces based on epochs/types/offset. - Add new trace RPC
trace_transaction
. - Use hex encoding for the returned bytes in trace-related RPCs.
- Add new fields
latestCheckpoint
,latestConfirmed
, andlatestState
incfx_getStatus
. - Improve some RPC error reporting.
Performance Optimization
- Reduce the memory usage for maintaining more snapshots with the configuration
additional_maintained_snapshot_count
.
Bug Fixes
- Fix a possible OOM error when a full node is catching up.
- Fix a possible OOM error in transaction pool when an archive node is catching up.
- Return correct
block_number
incfx_getStatus
. - Fix a bug that makes the configuration
mining_author
require extra quotes to use a CIP-37 base32 address. - Fix a bug that the block traces may be incorrect if the pivot chain switches frequently.