mirror of
https://github.com/status-im/nimbus-eth1.git
synced 2025-02-28 03:40:46 +00:00
0f89c1f901
* Update comments and colouring of example metrics display * Update comment for import/download serialisation flag details When importing starts while peers are actively downloading, the system tends to loose download peers, most probably due to high system activity. By this behaviour, there will be extra waiting delays for finding and connecting to new download peers. For this reason, importing and downloading is serialised. Downloading does not take place while importing. * Update comment on import start condition details When importing starts while peers are actively downloading, the system tends to loose download peers, most probably due to high system activity. By this behaviour, there will be extra waiting delays for finding and connecting to new download peers. For this reason, importing starts not before the staged blocks queue is filled up. The ramp up time to fill up is a fraction of the potential waiting time when losing peers. * Update comment on header or block fetch conversation via eth/XX * Increase staged blocks queue why Better overall throughput for slightly increased memory usage * Reduce header queue limits why In regular circumstances, the header queue has a few records most of time. Longer queues appear with unwieldy peers (bogus data, timeouts, etc.) if they happen to lock the lowest record so preventing from temporarily serialising the queue.