#evernode_update #evernode_changelog
Commit ID: c08e2a9a14bb137888a71ed89319e88202a35549
Node Version: 0.55.63
Tools version: 0.1.314
Block version: 44
📌 Update Instructions
1. Build node and node tools
2. Stop the node
3. Update
global config
4. Change node config (backup your config file before change):
config file: config.json
add
"states_cache_mode": "Moderate",
"skip_saving_persistent_states": false
check and change
"cells_db_config": {
"states_db_queue_len": 1000,
"max_pss_slowdown_mcs": 750,
"prefill_cells_counters": false,
"cache_cells_counters": true,
"cells_lru_size": 1000000
}
5. Start the node. Whait until sync_status will change from:
"node_status": "checking_db"
to
"sync_status": "synchronization_finished"
or "validation_status:" become Active
‼️ Attention: do not update the next node until the previous one has worked successfully for 3 hours.‼️
🛠Changelog
node from 0.55.35 to 0.55.63
🔹 Remove time limit for boot
🔹 Simplify procedure to get next key blocks for boot
🔹 Partially support hardforking in node
🔹 Produce applied shard hashes and raw block proofs to Kafka
🔹 Print WARN if node was built without --release
🔹 Backward compatibility for old and new broadcast algorithms
🔹 Speeding up saving persistant state
🔹 Speed up node shutdown
🔹 Fix for deep recursion on catchain blocks dropping
🔹 Limit for rocksdb log file: max 3 files, 100MB each
🔹 Improve internal message queue cleanup
🔹🔹 Add new algoritm for internal message queue cleanup with sorted access
🔹🔹 Previous algoritm with random access is optional (work if collator have free time)
🔹 Update metrics
🔹🔹 change cargo feature metrics to log_metrics, prometheus
🔹🔹 Support collect metrics to prometheus directly
🔹 Refactor processing of external messages
🔹🔹 external messages processing by FIFO (before was random access)
🔹🔹 fast track for external messages to non-existent account
🔹 Added shard states cache modes
🔹🔹 "Off" states are saved synchronously and not cached.
🔹🔹 "Moderate" (recommended), states are saved asiynchronously. Number of cached cells (in the state's BOCs) is minimal.
🔹🔹 "Full" (previous state), states saved asynchronously. The number of cells in memory is continuously growing.
🔹🔹 Added "cache_cells_counters" param in config.json. If it is set to true, then cell counters are cached in memory. It speeds up cells' DB but takes more memory. This cache was always enabled before.
🔹🔹 Added limit for cells DB cache size ("cells_lru_size" param in config.json). Zero value means default limit - 1000000.
evernode tool from 0.1.310 to 0.1.314
🔹 Support hops check for old-fashioned broadcasts
🔹 delete ever-crypto dependency
adnl from 0.8.19 to 0.9.1
🔹 Introduce server ID
🔹 ADNL channel reset on
overlay from 0.6.210 to 0.6.240
🔹 Overlay message sanity check
🔹 Allow to send duplicated broadcasts
🔹 Transparently support hops check for old-fashioned broadcasts
rldp from 0.7.205 to 0.7.232
🔹 Added RLPD message sanity check
ever_abi from 2.3.129 to 2.3.147
🔹 Use crypto functions from ever-types
ever_api from 0.3.24 to 0.3.41
🔹 Add raw.getShardAccountMeta
🔹 Add raw.getAccountByBlock
🔹 Add raw.getAccountMetaByBlock
🔹 Add raw.getAppliedShardsInfo
ever_block from 1.9.88 to 1.9.104
🔹 Added new method MerkleUdate::apply_for_with_metrics()
🔹 Fixed deserialization of BlockInfo. Added method BlockInfo::read_from_ex()
🔹 Fixed Block::set_gen_utime_ms method
ever_block_json from 0.7.168 to 0.7.187
🔹 Pruned cells are not serialized to BOC, only hash is written
ever-types from 2.0.18 to 2.0.28
🔹 Add BocWriterStack for faster boc saving
🔹 Transform BocWriter::traverse() into iterative algo
🔹 Refactor hashmap functions to optimize perfomance
ever_vm from 1.8.189 to 1.8.211
🔹 Add benchmarks
🔹 Recover load_hashed_cell() perf
🔹 Optimize IntegerData::as_slice()
🔹 Drop trees of StackItem::Tuple iteratively
Updated dependencies only
🔹 dht from 0.6.30 to 0.6.61
🔹 ever_executor from 1.16.58 to 1.16.82
ever-crypto (deprecated)
🔹 don't use in node packages anymore
No changes
🔹 lockfree
EV 