You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We loose the ability to restrict reth process on the os level. (threading / io)
Its no longer clear how many threads do we have and how can we schedule our work more efficiently because we fight with the reth internal jobs.
Lifetime of builder process is now tied to reth. Right now builder often kills itself to clean up some weird state. Now this approach will be sub-optimal because stopping reth prevents it from syncing and clears mempool / peers.
Mempool can fall out of sync on node restart
We can do something about it though but it will require some work and change of internals. (e.g. watchdog instead of killing the process will restart builder inside reth). We can use our own threadpools and never touch global ones (rayon).
also need to get playground working with it.
The text was updated successfully, but these errors were encountered:
Problems:
We can do something about it though but it will require some work and change of internals. (e.g. watchdog instead of killing the process will restart builder inside reth). We can use our own threadpools and never touch global ones (rayon).
also need to get playground working with it.
The text was updated successfully, but these errors were encountered: