Hodl
Hodl
I'm requesting them with full transactions and don't use redis cache. I just tested requests via ws (eth_getBlockByHash) with full transactions and it is fast as expected.
Yup, ws requests go through Dshackle. The current test is: I subscribed to newHeads and on update, immediately request eth_getBlockByHash with full transactions. Even If I restart Dschackle, results are...
Everything you stated is correct. getBlock is slow over http and almost no overhead over ws. And it's also true, that ws test is requesting only the latest block where...
You were right. The ws test was fast, because it fetched cached / latest block. Here are the results, http vs ws: ``` Block Number HTTP Benchmark (ms) WS Benchmark...
I also tried with redis and results are the same.
Is there any ETA for the fix? Or maybe some pointers how to properly fix the problem? I also noticed very high CPU load even when dshackle is idle (no...
Hi. Sorry for late feedback (i was on holidays). I applied the fix but results seem the same: ``` get full block: Direct node connection: 1. Current block: 17962783, block_size:...
I wrote simple python script: https://gist.github.com/tomih/177c40722b06a4de94aff9fa27376b70 Just change / add nodes to nodes variable. Let me know the results.