nodeos will not stay in sync on 2.0.0-rc2

[matthewdarwin]

When running nodeos 2.0.0-rc2 on EOS Mainnet on a public p2p endpoint with around 30 incoming connections, nodeos will sometimes not stay in sync. It sometimes falls behind by 1 or 2 minutes. After a (a few hours) it will catch up again. It is not predictable what is going on as it is accepting incoming connections from the general public.

Server is not too fast: Intel(R) Xeon(R) CPU E5-2690 0 @ 2.90GHz
Server has lots of RAM and is using the following configuration

wasm-runtime = eos-vm-jit
eos-vm-oc-enable = true
database-map-mode = locked

This problem was not experienced on 1.8.x. (Although of course EOS Mainnet transaction volume does change). Upgrade to 2.0.0 occurred on Nov 14.

The server is connected to other peers (locally on the same network) that are in sync.

Logs are available if required.

[sampleblood]

is this issue solved?

[matthewdarwin]

Work in progress. See #8229

[matthewdarwin]

The PR referenced above definitely helps.

[heifner]

Resolved by #8229