Makes it sound like we had a close call here… Was it really that bad?
Ah, thanks for the info. I was seeing the same behavior on the testnet. Makes sense now.
And it was probably discovered on the testnet first because people are more willing to go wild with submitting malformed transactions. Hopefully there’s not too many more of these bugs waiting to be discovered.
It was fixable even after the fact, but it could have frozen transactions on mainnet without a patch.
If it was fixable, it was also testable. That kind of implies that Chia Net is using farmers / pools as testers (less expensive than running their own QA, and this implies that more will come).
Do you know, when that bug was introduced? Was it introduced recently, or rather it is an old bug?
It looks like where all these WARNING add_spendbundle/Block validation time took their begining with. The question is still open: if i see these messages does it mean my chia softwere isn’t patched.
They tested it, found it, patched it and tested the patch in testnet. Not sure what else you are looking for?
it means your software IS patched these warnings you now see were INFO messages in previous builds (they still are if the validation time is < 1s)
Understand that chia team can’t test everything (if that were possible, apple would have enough resources to do so, and they’re loaded with bugs)
Also there have been much worse bugs in both bitcoin and ethereum, many times, of which ethereum had a much worse one less than 2 months ago (causing a chain split) … it just happens with highly complex software like chia is, especially when it’s still brand new (just 6 months on mainnet) … so younger software, and less severe bug → chia does a much better job than ethereum so far.
Who are these warnings for? Me? I’m not sure what they about and how to fix them.
You can ignore them …it just means your node took X time to validate a transaction block. It’s purely cpu speed times block size … it does become an issue when it’s constantly taking like more than 10 seconds but 1 or 2 s is fine
Usually it takes 2-3 secs, very rarely is 10+ secs. How could it be cpu issue if it’s pretty fast, I should not do plotting on node pc? Main question is can it affect winning probability?
The patch was implemented in 1.2.8. If you have installed 1.2.8 then you are patched.
Well, as you wrote in your article, XCHPool plus few other pools found and tested it. They let Chia Net know about it, and after that it was addressed in v1.2.8, or maybe in v1.2.9, as that was released soon after.
Glad to see that you brought up Bitcoin and Ethereum. Last time I checked, those were public open-source projects. As such code developed / tested / deployed by volunteers like you and me. Sometimes we do have time, sometimes we don’t; sometimes we do have the expertise, sometimes we don’t. Chia Net is a private company that got plenty of VC money also to hire right people to push the project.
Thanks for the replies, it looks like Chia’s communication could have been better.
In the end the issue would not have stalled mainnet, as there was a way to remove the bad TX from the mempool if it came to that.
It would have stalled the usability of mainnet, including pools claiming rewards. Block height would have continued to increase, but a full node still needs to put the reward transaction into a block for someone to receive it.