We have a parachain testnet running on a local relaychain based on rococo v2.
The Parachain is based on Cumulus (Polkadot and Substrate) v0.9.36 - so at the time of writing is the most up-to-date version;
We are attempting a large batch transaction making transfers from one user to a very large number (n=1000) of users - a simulation of a coin distribution after a crowdloan.
The transaction broke the parachain because the batch transaction apparently caused the block production to take took too long. The block time of 909ms mentioned seems absurdly short for it to break block production on a parachain that basically had a 12 second block time.
This also looks a lot like an attack vector, which would concern us greatly as we also have a production chain.
Is there any way to purge the extrinsics in the queue or recover the parachain some other way? It does not seem possible to restart the parachain once it has started producing blocks.
2023-01-13 18:11:18 [Parachain] ⌛️ Discarding proposal for slot 139469456; block production took too long 2023-01-13 18:12:54 [Parachain] 🎁 Prepared block for proposing at 563273 (909 ms) [hash: 0x2ad58e983e647f8f97dd955c0bfb0c0d7302465bc8ba1825184ce533f770073a; parent_hash: 0x790f…6483; extrinsics (3): [0xcd8e…bd1b, 0xb911…e88d, 0x85da…adb5]]