Node stays behind approximately distance to snapshot.

[trufa]

I have installed I'm running a base client pointed towards my up to date, synced L1.

eth_blockNumber of my L2 stays behind for approximately the same distance to the end of the snapshots, and then keeps the pace but doesn't' advance.

For example if I point HOST_DATA_DIR to the /reth-data (empty) it starts syncing very fast but from the very beginning

If I point it to my snapshot it starts also syncing very fast

node-1 | t=2024-10-26T01:58:38+0000 lvl=info msg="Advancing bq origin" origin=0x42bccb52f4bcdfd7cb77145632f14f98e3c276c8b370d0322c5a7ec68d683231:21015340 originBehind=false

but until certain point and they kind of keeps up but always lagging behind around 3-4 days which is the distance to the end of my snapshotted data.

I've tried running it --l1.trustrpc but doesn't seem to make any difference.

The following command as suggested by the docs:

echo Latest synced block behind by: $((($( date +%s )-
$( curl -s -d '{"id":0,"jsonrpc":"2.0","method":"optimism_syncStatus"}' -H "Content-Type: application/json" http://localhost:7545 |
jq -r .result.unsafe_l2.timestamp))/60)) minutes

Always keeps falling behind.

Any pointers would be greatly appreciated.

The errors I see in the log is one that I don't think is related

execution-1 | ts=2024-10-26T01:08:00.38459394Z level=error target=engine::jwt-validator message="Invalid JWT: Authorization header is missing or invalid"

And one that might be

node-1 | t=2024-10-26T01:51:23+0000 lvl=warn msg="Engine temporary error" err="temp: failed to sync forkchoice with engine: context deadline exceeded"

The peer count is in general around 8 and goes up to 11

[jun0tpyrc]

i was/am still hitting same thing , can you check log of reth if synchronising in batch mode?!

OP_NODE_SYNCMODE=execution-layer
OP_NODE_L2_ENGINE_KIND=reth <== seem without this  https://docs.optimism.io/builders/node-operators/configuration/consensus-config#l2enginekind reth doesn't go into batch mode for me, and it is missing in env example

and i also needed to tune this to be 10x larger https://reth.rs/run/config.html#merkle

then i can go to in-sync from similar scenario of few days, but still some performance issues unsolved occasionally(random beteween few seconds - 1hour delay some times ) on reth (already on nvme & not seeing obvious cpu bottleneck)

[trufa]

@jun0tpyrc I honestly gave up trying to make reth work and now switched to geth, it seems to be syncing reasonably fast now.

[lacotWin]

I had quite hard times here. I experienced static 12h drag and struggled a lot to fix it.
OP_NODE_L2_ENGINE_KIND=reth
fixes the problem for me. I believe it should be documented in this guide as there is no mention of it, not even in .env files

[jun0tpyrc]

I had quite hard times here. I experienced static 12h drag and struggled a lot to fix it. OP_NODE_L2_ENGINE_KIND=reth fixes the problem for me. I believe it should be documented in this guide as there is no mention of it, not even in .env files

wonder what spec you on &/ able to stay within minute of delay on reth for base-mainnet?! 🙏

[azeonmarket]

peer count 8-11 is quite low for an optimized performance, but if you managed to fix the issue great to hear ! true that OP_NODE_L2_ENGINE_KIND=reth is something you might have to add on configs, but its something you dont necessarily need but can be useful to know if you encounter some problems.