Alex Sharov
Alex Sharov
Erigon doesn't store traces in db. OE does. We don’t plan to change it in near future (low priority task).
1. I think it's because our StageBodies is not limited by size (and downloading of blocks in background is also not limited) 2. add --txpool.disable to reduce scope
3. if add --pprof: then can do things like go tool pprof -lines -http=: http://127.0.0.1:6060/debug/pprof/heap
show recent Erigon's logs
mdbx_stat - must be fast. try to stop erigon.
`GC: 38621002 2.4%` = 150Gb - quite big number - probably it causing `Commit cycle in=33.843631406s` It will get better in 1-3 days. Why it happend: - May be related...
OS? Do you run mdbx_stat from inside container or from outside?
Prometheus/Grafana - no Geth attach - i don’t remember how it works under the hood - probably websocket subscription - then also must be not a reason Your custom script...
Check that `GC: 38621002 2.4%` section of mdb_stat reducing with time. If yes - then it will get better and better. But you have no obvious reasons - why it...
Not related