Low sync committee participation using dev command

[wemeetagain]

Describe the bug Running a transient devnet using the dev command, sync committee participation is very low (often zero).

Expected behavior

Using the dev command, the sync committee participation should be perfect.

Steps to Reproduce

node --inspect lodestar dev --api.rest.host 0.0.0.0 --api.rest.api beacon config events node validator lightclient --startValidators 0:7 --params.ALTAIR_FORK_EPOCH 0 --params.SECONDS_PER_SLOT 6 --sync.isSingleNode --network.discv5.enabled false --logLevel verbose --logLevelFile debug --logFile foo.log

Separately:

curl "http://localhost:9596/eth/v2/beacon/blocks/N" | jq .

[dadepo]

From what I can see, it seems the sync committee participation is now ok. Given the following output:

➜  ~ curl "http://localhost:9596/eth/v1/beacon/states/head/sync_committees" | jq .
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100   179  100   179    0     0  17640      0 --:--:-- --:--:-- --:--:-- 44750
{
  "data": {
    "validators": [
      "3",
      "2",
      "5",
      "6",
      "7",
      "4",
      "0",
      "1",
      "3",
      "2",
      "5",
      "6",
      "7",
      "4",
      "0",
      "1",
      "3",
      "2",
      "5",
      "6",
      "7",
      "4",
      "0",
      "1",
      "3",
      "2",
      "5",
      "6",
      "7",
      "4",
      "0",
      "1"
    ],
    "validator_aggregates": []
  }
}

also

curl "http://localhost:9596/eth/v2/beacon/blocks/2" | jq -r '.data.message.body.sync_aggregate'
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  1883  100  1883    0     0   315k      0 --:--:-- --:--:-- --:--:--  919k
{
  "sync_committee_bits": "0xffffffff",
  "sync_committee_signature": "0xaa5701bb45a9733fc52967f9f6fcd5f5ce378a1b61293031360d6895b51184a3febe742938c99039ab46dfc5dd45d1b40451f8c2b9b65587879f46491cac5e893c6dbf7fd16aa79515a6e2b604e165653d1e1e3dcc63193373afc52de7cf439b"
}

Which I interpret to mean that sync committees participated in producing a signature.

@wemeetagain Got any other ideas to further check if this is still an issue?

[philknows]

Sim test assertions being completed in https://github.com/ChainSafe/lodestar/issues/4276 should prove or disprove this issue.

[dapplion]

Old issue