Option A: Cosmos Hub 2 and Cosmos Hub 3 should not be running simultaneously. It will be a mess if Cosmos Hub 2 still running when Cosmos Hub 3 is online. This is not an option.
Option B: Testnet is always available to test new software features. Unless there is new major software release, we don’t have to bother new testnet. I agree we need an upgrade simulation.
Option C: To simulate the Cosmos Hub upgrade, @bharvest is suggesting to fork Cosmos Hub 2, start a new network with another chain-id. This is purely a simulation/practice of the upgrade process. If you would like to use your testing node to simulate this upgrade process, you can simply replace the validator key in your testing node to run this. A single validator running multiple validator nodes can be easily achieved. I think we already have enough keys to start this simulation. I don’t see why this is an unreal scenario. If this is smooth, I think we should make this as a best practice for future upgrade.
Option A : IMO is the best option and can be done. If can’t be done, then someone from Tendermint or Cosmos team can confirm I’m Wrong, and says this option CANT be done. But I’m very sure, like 99.99% that can be done.
and when= Option C : To simulate the Cosmos Hub upgrade, @bharvest is suggesting to fork Cosmos Hub 2.
here (This is not a fork neither) this is just one private simulation. not a fork.
BUT hey, here
If you want to do the simulation or others are planning or what someone wants to do, do it, I am not saying that you or others do not do it, I am saying that I will not participate because I do not see it.
in the same way here I am free to say why I do not participate and which in my opinion is the best option.
ant just to finish.
Im not sure if people understand when is a fork and when is not. I have that part very clear.
Problems on using gaia-13006 is that it is already based on new gaia version so we cannot test migrate function, which transforms genesis from old gaia to new gaia version.
Not a fork to be a mainnet
When we launch hub3, it should be based on proper governance procedure, so, whatever we launch from hub2, it is not a mainnet because it is not based on any onchain governance. Because it is not a mainnet, we should not use cosmoshub-3 as a chain-id.
more real environment testing
The mainnet ledger is the best environment to test the program because the ultimate target of the new software is mainnet. It will give us the best confidence.
Based on above 3 reasons, I think it explains why our approach is reasonable. But, it is not a mandatory to participate. It is a community self-motivated testnet who are not satisfied with last upgrade procedure, so that who wants to make things much more certain by our hands.
Just clarify
When I talk about kill 13006 is for start a new testnet 13007 soft 0.34.6.
That is very clear. And for that is testnet. For public testnet, and test software upgrade before do it on mainnet. Not sure the point of keep running 13006 and start your private testnet.
But like from 100 validators and core dev team and more people, Im the only one talking about this, just ignore me.