Node stops responding and memory use keeps growing

[ermalkaleci]

Trying to deploy flow-protocol-ethereum makes the node to increase memory use and not responding to rpc calls or producing new blocks. It will run fine for a while then it's like it gets stuck into a loop.

Steps to reproduce the issue

Start frontier node

1. Build specs

cargo run -- build-spec --dev > specs.json

2. Update specs.json to give 0xb393de29a6a341a32758735a160f3561b88fded0 some balance

"frameEvm": {
  "accounts": {
    "0xb393de29a6a341a32758735a160f3561b88fded0": {
      "nonce": "0x0",
      "balance": "0xffffffffffffffffffffffffffffffff",
      "storage": {},
      "code": []
    }
  }

3. Start frontier node

cargo run --  --chain=specs.json --execution=Native --validator --alice --force-authoring --rpc-port=8545 --tmp

Deploy flow-protocol-ethereum

git clone git@github.com:laminar-protocol/flow-protocol-ethereum.git
cd flow-protocol-ethereum
git checkout frontier
yarn install
yarn deploy:development

the deployment will go well for a while and after a few minutes the node will stop responding and producing blocks and memory usage will keep growing.

[sorpaas]

If the contract does not use CREATE2, then evm v0.18.3 should fix everything even in its stack runner. Contracts with CREATE2 might work but I'm still debugging one final issue.