Multiple Node Production Chain occupies too much memory

[Geo0426]

Blockchain occupies too much memory

Does parameter dev_ block_ interval_ ms only take effect in dev mode?

I chose the test mode, but I don't know how to reduce the block out time. My server has limited memory, and I don't know the reason for the high memory usage. I guess it's because the mining speed is too fast (possibly).

What should I do to reduce memory usage?

Sorry, my current server resources are limited. So I want to know how I can reduce the memory usage of independent chain. Should I modify the configuration file?

conflux.log

2023-04-04T10:03:28.824010219+08:00 WARN IO Worker #2 network::ser - No peers connected at this moment, 1 pending + 0 started
2023-04-04T10:03:28.824049441+08:00 INFO IO Worker #1 cfxcore::syn - Catch-up mode: false, latest epoch: 407 missing_bodies: 0
2023-04-04T10:03:28.902911961+08:00 INFO mining cfxcore::syn - Mined block 0xc82f2783c3207a379cfa9b6219e2748880430f1e21f745ce45eb847f27e55e20 header=BlockHeader { rlp_part: BlockHeaderRlpPart { parent_hash: 0xfe104eaa03a4c5da92f77461218c97f3bcb6ebfee39ba37390015c5f299f4167, height: 408, timestamp: 1680573808, author: 0x174b46fc33fcd702db97871183a882e1f9172f94, transactions_root: 0xc5d2460186f7233c927e7db2dcc703c0e500b653ca82273b7bfad8045d85a470, deferred_state_root: 0xae20b2ede475b5c72490f995adf15a5b19670d138cea7dc3404c7b2064163c08, deferred_receipts_root: 0x09f8709ea9f344a810811a373b30861568f5686e649d6177fd92ea2db7477508, deferred_logs_bloom_hash: 0xd397b3b043d87fcd6fad1291ff0bfd16401c274896d8c63a923727f077b8e0b5, blame: 0, difficulty: 4, adaptive: false, gas_limit: 30000000, referee_hashes: [], custom: [], nonce: 15116217592705301801 }, hash: Some(0xc82f2783c3207a379cfa9b6219e2748880430f1e21f745ce45eb847f27e55e20), pow_hash: None, approximated_rlp_size: 352 }
2023-04-04T10:03:28.993693415+08:00 INFO mining cfxcore::syn - Mined block 0xa0f68944d8aadca8d21bd24a8047553d4de64df481187fb675974023f03f1943 header=BlockHeader { rlp_part: BlockHeaderRlpPart { parent_hash: 0xc82f2783c3207a379cfa9b6219e2748880430f1e21f745ce45eb847f27e55e20, height: 409, timestamp: 1680573808, author: 0x174b46fc33fcd702db97871183a882e1f9172f94, transactions_root: 0xc5d2460186f7233c927e7db2dcc703c0e500b653ca82273b7bfad8045d85a470, deferred_state_root: 0x40d422654c680af813ce0cebadcd2d32a847a1fa79f506ee5bc3d3ea408692ee, deferred_receipts_root: 0x09f8709ea9f344a810811a373b30861568f5686e649d6177fd92ea2db7477508, deferred_logs_bloom_hash: 0xd397b3b043d87fcd6fad1291ff0bfd16401c274896d8c63a923727f077b8e0b5, blame: 0, difficulty: 4, adaptive: false, gas_limit: 30000000, referee_hashes: [], custom: [], nonce: 6398255997441759358 }, hash: Some(0xa0f68944d8aadca8d21bd24a8047553d4de64df481187fb675974023f03f1943), pow_hash: None, approximated_rlp_size: 352 }
2023-04-04T10:03:29.114341123+08:00 INFO mining cfxcore::syn - Mined block 0xc375d8c6c1fd28716a3ea53d69d350eab1af0c4347888587e5834152a6407b5a header=BlockHeader { rlp_part: BlockHeaderRlpPart { parent_hash: 0xa0f68944d8aadca8d21bd24a8047553d4de64df481187fb675974023f03f1943, height: 410, timestamp: 1680573809, author: 0x174b46fc33fcd702db97871183a882e1f9172f94, transactions_root: 0xc5d2460186f7233c927e7db2dcc703c0e500b653ca82273b7bfad8045d85a470, deferred_state_root: 0x9d8f3d8eef85d03d17ece7aa739ab9c4af55b7bc7f4f85ce65614293585436a9, deferred_receipts_root: 0x09f8709ea9f344a810811a373b30861568f5686e649d6177fd92ea2db7477508, deferred_logs_bloom_hash: 0xd397b3b043d87fcd6fad1291ff0bfd16401c274896d8c63a923727f077b8e0b5, blame: 0, difficulty: 4, adaptive: false, gas_limit: 30000000, referee_hashes: [], custom: [], nonce: 11452726266917030730 }, hash: Some(0xc375d8c6c1fd28716a3ea53d69d350eab1af0c4347888587e5834152a6407b5a), pow_hash: None, approximated_rlp_size: 352 }
2023-04-04T10:03:29.205182073+08:00 INFO mining cfxcore::syn - Mined block 0xcb19c6dddf640cf5b3d74c0023934ee20a775c57f0133b9b3282204092fada25 header=BlockHeader { rlp_part: BlockHeaderRlpPart { parent_hash: 0xc375d8c6c1fd28716a3ea53d69d350eab1af0c4347888587e5834152a6407b5a, height: 411, timestamp: 1680573809, author: 0x174b46fc33fcd702db97871183a882e1f9172f94, transactions_root: 0xc5d2460186f7233c927e7db2dcc703c0e500b653ca82273b7bfad8045d85a470, deferred_state_root: 0x31bba6fe1fc66b8b4e081c2e5faf859d79bfcefbe4a518989277a0ee67527c3f, deferred_receipts_root: 0x09f8709ea9f344a810811a373b30861568f5686e649d6177fd92ea2db7477508, deferred_logs_bloom_hash: 0xd397b3b043d87fcd6fad1291ff0bfd16401c274896d8c63a923727f077b8e0b5, blame: 0, difficulty: 4, adaptive: false, gas_limit: 30000000, referee_hashes: [], custom: [], nonce: 5754150310821372984 }, hash: Some(0xcb19c6dddf640cf5b3d74c0023934ee20a775c57f0133b9b3282204092fada25), pow_hash: None, approximated_rlp_size: 352 }

[peilun-conflux]

1. dev_block_interval_ms only takes effects in dev mode where blocks are generated with fixed interval for a single server.

2. By default a single node takes up to 13GB memory and the suggested memory requirement is 16GB. To reduce the memory usage, you could adjust the following performance-related cache configurations: ledger_cache_size, rocksdb_cache_size, and several storage_delta_mpts_cache_*_size. You can try to reduce them all by the same ratio to see if the memory usage changes. Other consensus-related configurations dev_snapshot_epoch_count (2000) and era_epoch_count (20000) could also be reduced, but this may lead to unexpected crash if not configured properly.