Jacek Sieka
Jacek Sieka
When migrating from another client, our migration guide should suggest how to trusted-node-sync from the source client, to avoid a long way - this way, client migration can be completed...
When something is wrong with eth1 monitoring, for example the EL is syncing, we print an unintelligible parse error: ``` "Eth1 chain monitoring failure, restarting topics="eth1" err="getBlockByHash(m.dataProvider,\n BlockHash(m.depositsChain.finalizedBlockHash.data)) failed 3...
Trying to fetch deposits on an unsynced EL results in addtional load on the EL further delaying its sync and leads to timeouts in the API - it's also bound...
No "Slot start" message is printed, thus it's hard to tell if the light client is doing anything, and what its latest view of the network is (what it thinks...
shutting down the lc gives an ugly message ``` ^CTraceback (most recent call last, using override) /home/arnetheduck/status/nimbus-eth2/vendor/nim-testutils/testutils/moduletests.nim(21) moduletests /home/arnetheduck/status/nimbus-eth2/beacon_chain/nimbus_light_client.nim(236) main /home/arnetheduck/status/nimbus-eth2/vendor/nim-chronos/chronos/asyncloop.nim(710) poll /home/arnetheduck/status/nimbus-eth2/vendor/nim-chronos/chronos/ioselects/ioselectors_epoll.nim(379) selectInto /home/arnetheduck/status/nimbus-eth2/vendor/nimbus-build-system/vendor/Nim/lib/system/excpt.nim(610) signalHandler SIGINT: Interrupted by Ctrl-C....
So that it doesn't have to be supplied on every restart, we should persist basic light client information such as the current head / finalized checkpoint, so that it can...
Each duty is being received twice for some reason ``` INF 2022-08-18 21:03:36.042+02:00 Received new sync committee duty duty="(pubkey: 8c5e31e2c0c101c4e48b08cf6a654b4d55e997c8b87d207f585783fd9b37077866941106315f02ebd38de04ea1b8962b, validator_index: 212563, validator_sync_committee_index: 128)" epoch=115472 service=duties_service INF 2022-08-18 21:03:36.124+02:00 Received...
When we're opt-synced, we already know the duties for the coming slots - we can use them to print when the next attestation / block will happen. We can also...
When the connection to the web3 server is lost, it remains possible and valid to keep voting for the last known head for some time (1 epoch? up to 2?)...
When we are only optimistically synced, we should not be attesting / producing blocks: https://github.com/ethereum/consensus-specs/blob/dev/sync/optimistic.md#validator-assignments Here's a log from a time when the beacon node is synced but the underlying...