Testing sync #650
Replies: 4 comments 2 replies
-
so you're getting incoming connections? The useless peer is then emitted by the remote peer, because we currently don't send this @Rjected |
Beta Was this translation helpful? Give feedback.
-
is there a chance you are setting the debug tip incorrectly? the hash of block 16289962 is
|
Beta Was this translation helpful? Give feedback.
-
@jonasbostoen thank you for |
Beta Was this translation helpful? Give feedback.
-
I have a synced Geth instance in the cloud. We can pair our nodes if you wish. Peace |
Beta Was this translation helpful? Give feedback.
-
This is a discussion in which I'll post the findings on trying to sync a node.
Context
I ran the node for 30 minutes trying to sync up to block
16289962
with the following command:During this time, it wasn't able to import any headers from the p2p network. This is the state of the db at the end:
So only the genesis header was written.
The debug logs can be found in this TG message: https://t.me/paradigm_reth/939
Errors
P2P
Looking at the p2p errors we get the following:
UselessPeer
errors (of which 763 on inbound connections)TooManyPeers
errorsIt was hard to connect to peers, got up to around 8 at the end, but they were useless since reth wasn't requesting any headers from them.
Header stage
Every once in a while we connected to a peer and started requesting headers, but we never really got any. Usually we got
EmptyResponse
errors:Or just 0 len headers:
At one point I also got a
TheMergeMixHashIsNotZero
error:This originates here:
reth/crates/consensus/src/verification.rs
Lines 60 to 61 in de6630a
This was probably a bad block, can't find the hash anywhere.
Next Steps
I think it might be useful to try to connect
reth
to a syncedgeth
instance, which is what I'm going to do next.Beta Was this translation helpful? Give feedback.
All reactions