liftedinit / manifest-ledger

CosmosSDK-based blockchain ledger for the Manifest Network
Apache License 2.0
0 stars 2 forks source link

Unable to start a new validator #30

Closed fmorency closed 2 months ago

fmorency commented 3 months ago

The issue is two-fold

  1. Unable to build tag v0.0.1-alpha.1 because of

    go: github.com/reecepbcups/tokenfactory@v0.50.0-alpha.3: reading github.com/reecepbcups/tokenfactory/go.mod at revision v0.50.0-alpha.3: unknown revision v0.50.0-alpha.3

    unless you already have the package in your cache.

  2. Unable to start a new validator using the main branch because of

    panic: invalid ibc channel genesis state parameters: upgrade timeout timestamp invalid: 0: upgrade timeout is invalid [cosmos/ibc-go/v8@v8.1.0/modules/core/04-channel/types/params.go:34]

The following instructions were used to create and start the new validator

git clone https://github.com/liftedinit/manifest-ledger.git
cd manifest-ledger
git checkout v0.0.1-alpha.1  # Also tried with the `main` branch, see above
make install

manifestd init <MONIKER> --home ~/.manifest --chain-id obvious-1
cd ~/.manifest/config
mv genesis.json genesis.json.orig

wget https://raw.githubusercontent.com/liftedinit/manifest-ledger/bd75ab261749d2eea8d50e765a08237e8aafa6ad/network/obvious-1/live_genesis.json
mv live_genesis.json genesis.json

# Edit genesis.json and add the following to `persistent_peers`
persistent_peers = "cf641df39cce560d43c9889ff0883772de2e8a59@15.204.143.232:26656,dddb7b8ad73be43f3dc96922d6708d53e847685d@66.172.36.142:21156"

manifestd start

CC @jgryffindor

chalabi2 commented 3 months ago

Should be fixed in v0.0.1-alpha.3 Latest

fmorency commented 2 months ago

@chalabi2 We cannot start a node with alpha.3 against the current testnet. @jgryffindor and I are getting the same error as before.

fmorency commented 2 months ago

@chalabi2 @Reecepbcups

Using

https://github.com/liftedinit/manifest-ledger/blob/reece/test-network-launch/network/obvious-1/live_genesis.json

and building off

https://github.com/liftedinit/manifest-ledger/tree/reece/test-network-launch

I was able to start a node successfully. However, I got the following after a while

1:41PM INF Reconnecting to peer addr=dddb7b8ad73be43f3dc96922d6708d53e847685d@66.172.36.142:21156 module=p2p
1:41PM INF service start impl="Peer{MConn{66.172.36.142:21156} dddb7b8ad73be43f3dc96922d6708d53e847685d out}" module=p2p msg="Starting Peer service" peer=dddb7b8ad73be43f3dc96922d6708d53e847685d@66.172.36.142:21156
1:41PM INF service start impl=MConn{66.172.36.142:21156} module=p2p msg="Starting MConnection service" peer=dddb7b8ad73be43f3dc96922d6708d53e847685d@66.172.36.142:21156
1:41PM ERR failed to add block err="requester is different or block already exists (peer: dddb7b8ad73be43f3dc96922d6708d53e847685d, requester: dddb7b8ad73be43f3dc96922d6708d53e847685d, block height: 192434)" module=blocksync
1:41PM ERR Stopping peer for error err="error with peer dddb7b8ad73be43f3dc9
...
1:42PM ERR Error in validation err="wrong Block.Header.AppHash.  Expected 972C57BFE51933E20458CE41C9182FF2AFF341FD260C387071D826998CFA7339, got DD49740C30E6874F9FA79D677C8C769DDC9046AA5CF90D7D2CD405B6C67641BE" mod
ule=blocksync           

Something seems to be breaking consensus.

Reecepbcups commented 2 months ago

@fmorency did you start the network with POA_ADMIN_ADDRESS=manifest1nzpct7tq52rckgnvr55e2m0kmyr0asdrgayq9p manifestd start where POA_ADMIN_ADDRESS is an env variable? (as described in the testnet branch summary)

I assume 192434 is where I ran a POA admin command, which if the POA_ADMIN_ADDRESS was not set properly, the node would have faulted there. THis only is required for testnet due to some SDK work arounds.

You can manifestd rollback (or manifestd rollback --hard if the first does not work) and start again

fmorency commented 2 months ago

@fmorency did you start the network with POA_ADMIN_ADDRESS=manifest1nzpct7tq52rckgnvr55e2m0kmyr0asdrgayq9p manifestd start where POA_ADMIN_ADDRESS is an env variable? (as described in the testnet branch summary)

I assume 192434 is where I ran a POA admin command, which if the POA_ADMIN_ADDRESS was not set properly, the node would have faulted there. THis only is required for testnet due to some SDK work arounds.

You can manifestd rollback (or manifestd rollback --hard if the first does not work) and start again

That fixed it! Thank you! 🙏🏻