node sync problem

[shenyaqi9527]

My nodes are slow to synchronize. After syncing to the latest block, after a while, it is behind the latest block.
my service is 4c16G. Do I need to upgrade my service?

INFO [08-01|10:29:18.706] Imported new chain segment               blocks=1 txs=18  mgas=13.863  elapsed=26.078s   mgasps=0.532  number=3779491 hash=f46ed2…00cd27                                                      cache=268.33mB
INFO [08-01|10:29:31.647] Imported new chain segment               blocks=1 txs=8   mgas=7.949   elapsed=12.940s   mgasps=0.614  number=3779492 hash=a3537e…90ce5d                                                      cache=268.33mB
INFO [08-01|10:30:01.687] Imported new chain segment               blocks=1 txs=28  mgas=9.325   elapsed=30.040s   mgasps=0.310  number=3779493 hash=c19397…070257                                                      cache=268.33mB
INFO [08-01|10:30:30.420] Imported new chain segment               blocks=1 txs=13  mgas=8.651   elapsed=28.732s   mgasps=0.301  number=3779494 hash=445bad…f02ec6                                                      cache=268.33mB
INFO [08-01|10:30:59.189] Imported new chain segment               blocks=1 txs=18  mgas=10.260  elapsed=28.768s   mgasps=0.357  number=3779495 hash=2d3daf…c016c9                                                      cache=268.33mB
INFO [08-01|10:31:57.424] Imported new chain segment               blocks=2 txs=35  mgas=20.570  elapsed=58.235s   mgasps=0.353  number=3779497 hash=6c01e8…d733c9                                                      cache=268.33mB
INFO [08-01|10:32:09.651] Imported new chain segment               blocks=1 txs=11  mgas=1.307   elapsed=12.227s   mgasps=0.107  number=3779498 hash=354629…6801ba                                                      cache=268.33mB
INFO [08-01|10:33:22.908] Imported new chain segment               blocks=4 txs=42  mgas=24.536  elapsed=1m13.257s mgasps=0.335  number=3779502 hash=74e563…fabb4d                                                      cache=268.33mB
INFO [08-01|10:33:46.929] Imported new chain segment               blocks=1 txs=14  mgas=3.959   elapsed=24.021s   mgasps=0.165  number=3779503 hash=8aa9b8…19188d                                                      cache=268.33mB
INFO [08-01|10:34:03.385] Imported new chain segment               blocks=1 txs=9   mgas=2.819   elapsed=16.455s   mgasps=0.171  number=3779504 hash=438a45…370bad                                                      cache=268.33mB

[kriswebdev]

Node sync is inefficient when trying to download the full chain from scratch.

Gero bottleneck during node sync is disk I/O, not network.

~1000 files are modified per minute in chaindata folder (Google leveldb), resulting is disk use (SSD) almost 100%. Node sync speed decreases drastically over time:

• ~500 blocks/min at block 2.400.00 (chaindata: 42GB)
• ~70 blocks/min at block 3.200.000 (chaindata: 93GB)
• We are currently at block 7.000.000! It would take weeks to sync if the rate doesn't decrease again!

There is a big scalability issue.

There should be a different way to download / verify / store the chain data when the client is very late behind the last block.

[DukeNukem-dev]

I would add that besides disk I/O, the CPU can also be a bottleneck. Recommend AMD Ryzen 3900x or higher. The constant disk IO requires many CPU threads and larger L3 cache.