How Ethereum’s Istanbul testing got delayed

The first topic from the Ethereum Devs meeting today addressed the Istanbul testnet delay. Péter Szilágyi spoke about how the upgrade happened two days earlier which caused people to think that the fork didn’t succeed. It turned out that a miner with around 50 GPUs was pushing the non forked chain. 

“The Ropsten fork was a bit of a shitshow,” said Szilágyi.

Istanbul was supposed to begin testing on the Ropsten block on October 2nd. 

What does it mean for the Ropsten testnet? 

It is too late to fix the situation. The only way for the Istanbul upgrade to progress on Ropsten is for more people to be mining the upgraded version than the non-upgraded one. 

Szilágyi said the team will be looking into ways to prevent this from happening again, but fixes will have to be implemented preemptively. 

It was also mentioned that there being a heavier load on the non-upgraded nodes makes it harder for people to start mining on the upgraded version. 

Tim Beiko said that all the other testnets have been performing as expected 

How the devs are reacting

Ethereum developers believe that the occurrence was valuable. Unfortunately, many of blockchain projects are in the process of learning as they go. 

Martin Holst Swende said that there hasn’t been any complaining about the Istanbul delay from DApp devs. Apparently, the last time they worked with Ropsten, it caused some problems for developers. 

A few proposals are in the work for Ethereum to handle these scenarios better. Swende says these forks bring a lot of stuff up that is good to know about

It may be a bit surprising to learn that they are kind of admitting having not considered what to do if a guy with 50 GPUs started mining on a testnet a bit early. The good thing is that there are three other blocks where Istanbul will be tested. Testing on the Görli block is scheduled to start October 30th.