Try   HackMD

Namada SE 2 Restart 1

There was a bug that was introduced in version v0.31.1 by accident. The bug involved a line mistakenly moved to an async section in Namada's ABCI interface code, which could have led to reordered ABCI events and explains much of the behaviour observed on the testnet.

v0.31.2 has been released, which fixes the bug. You can install either from source or use the binaries build in the CI.

ABOVE ALL ELSE, DON'T DELETE THE KEYS OR THE VALIDATOR STATE.

I would suggest the following steps to upgrade:

  1. Stop the validator node
  2. Move your <chain_dir> directory out of the <base_dir> by moving it somewhere else.
  3. Run join network again with join-network (the base directory must be empty) (this is exactly the same process as when starting)
  4. Modify the seeds and persisent peers (possibly grab others from Discord).
    • vim <base_dir>/shielded-expedition.88f17d1d14/config.toml
      • seeds = "20e1000e88125698264454a884812746c2eb4807@testnet-seeds.lavenderfive.com:20056,tcp://3c12dd31adf945bd10d551713d8dc22795b7de0c@194.163.172.168:26656"
      • persistent_peers = "tcp://dc9f803d6a0cbe6c84d20a6d8e4d4e138ccc9635@195.3.222.156:26657,tcp://5890efc822a0fa02d289c6e47cd2ee6b9a4ac693@62.171.156.190:26656,3c12dd31adf945bd10d551713d8dc22795b7de0c@194.163.172.168:26656,tcp://bbde1b8adbb136b743018c7f92f5ee0162433fb6@65.109.39.125:26656,0707d78aca8514a48fffc09803c31e59e7f495ed@34.16.191.152:26656,195763c651dc4586e95beaf11b19346eb4d907fe@34.28.125.251:26656,4ff683191183da06859f4987733803838f69f8f2@218.155.161.169:30600,4f7276ddcb8abe1bcfc92454f925fbe59b819895@34.16.197.7:26656,6553e277ae9cd9098ee6bc518328427eea69eb5b@45.33.86.221:26656,0546acfa046274f76d1f77b72c567768bd2d5cca@185.172.191.9:26656,2c17c1914d5f21b366149caa33bf81e3b2037ba5@34.163.114.26:26656,10f764862481d7545323a4537da711dce111103e@34.155.116.21:26656,672d139bbbea0ec0370bc468529a8f882d5ce862@34.163.230.173:26656,e522f80c98a3997da1701e9e241e8405b17e90fc@34.38.40.86:26656,e1be9c2d3bfee05353f3f213742f3f03edbd3096@34.163.229.97:26656,89300a5836b15315ee4c7f70591a3dc9946fff5d@52.19.225.58:26656,1c134003c7aff4a266f2d7078bc19ceaf2554adf@34.32.50.165:26656,d4e7227050d08b12b0f11526dc1c51d2527cf973@34.32.54.215:26656,b6ffac4f7fcbfaff478e77884d995507a1cf4e1a@35.205.168.17:26656,5f4ed5790c2a0465eb14db751c2c64a2d0638605@34.32.46.90:26656,75513794710941cd05d6e16753959d84c9d11dc7@34.118.44.6:26656,0a9567cbe74911bcd8883b65e5f904ed487785fc@34.154.58.111:26656,b62a83c1bdddaa8e9bef7b1509a58bcab030245b@34.118.26.153:26656,3b15b3844da042f793ac160525ebfb8b79a292f1@136.243.40.38:26656,e3d260468a72078b46d32adb7175ee4b95e2ced2@5.78.77.122:26656"
  5. Modify the consensus timeouts in order to speed up the sync
    • vim <base_dir>/shielded-expedition.88f17d1d14/config.toml
      • timeout_propose = "300ms" (used to be "3000ms")
      • timeout_propose_delta = "0ms" (used to be "500ms")
      • timeout_prevote = "100ms" (used to be "1000ms")
      • timeout_prevote_delta = "0ms" (used to be "500ms")
      • timeout_precommit = "100ms" (used to be "1000ms")
      • timeout_precommit_delta = "0ms" (used to be "500ms")
      • timeout_commit = "1000ms" (used to be "10000ms")
  6. Possibly replace the cometbft/data/priv_validator_state.json with the values from the previous directory. (this is up to your judgement and depends on your double-sign prevention setup)
  7. Download/Install from source 0.31.2, swap them for the old ones and start your node with namadan ledger run (adjust as needed)
  8. After the node as fully synced until block height and the latest round height (currently 500 but check Discord). Stop the node, revert the consensus timeouts to their original values and start the node again.

Notes

  • If everyone does it at the same time, then no one can serve the data, so please back-up your base_dir and potentially serve it over some other node.
  • No need to change the .masp-params.