The fault, which has already been fixed, had multiple factors that caused it.
Such a problem could not happen in Ethereum 2.0, which is more advanced in its development.
The testnet The Ethereum 2.0 Ropsten testnet had operational issues on Thursday, May 26. What happened was that the increase in the hashrate or computing power of the miners increased to such an extent that the terminal difficulty bomb (TDD) was activated prematurely, which precisely ends the proof of work.
A) Yes, should have happened what will happen on the Ethereum mainnet when the merger goes through. This is that, after the difficulty bomb is activated, the transactions begin to be validated in the Beacon Chain, that is, the original fragment of Ethereum 2.0 that will already work with proof of participation, Proof-of-Stake or simply PoS.
Nevertheless, the problem in Ropsten was that the own Beacon Chain of this testnet was not created yet. In fact, this is projected to happen on May 30, days before the testnet merger takes place on June 8.
Ropsten will be the first testnet to “test” merging from one consensus algorithm (PoW) to another (PoS). Then they will be followed by testnet Goerli and Sepolia.
What happened to Ropsten and how is the network now?
As a consequence of the inconvenience that was explained in detail by the developer Tim Beiko and what was it replicated by Twitter user @functi0nZer0, Ropsten was detained for several hours. When it resumed operation, the network processed blocks of a single transaction each, with the message Will recover ASAP (“It will recover soon”) in its metadata.
According to what the block explorer shows etherscan, the arrest occurred in block 12,307,424, around time 18 (UTC) on Thursday, May 26. Half an hour later, blocks with a single transaction began to be created, as mentioned. The next block with more than one transaction was 12,308,894, 10 hours after the testnet stopped.
More details on the failure in Ropsten
Not only does a Beacon Chain need to be created for the Ropsten merge to take place, but something else is also needed. We talk about the hard fork either hard fork Bellatrix, which will be the one to monitor the difficulty bomb that will end mining. It is scheduled for June 3.
The difficulty bomb is not calculated in time, but in a difficulty value that must be reached for it to activate. What Bellatrix will do is trigger the switch from one chain with PoW to another with PoS when that value is reached.
Precisely what happened with Ropsten and led to the network failure is that the increase in hashrate by a miner (no less than 30 times more than the previous hashrate) made the network difficulty too high. This setting occurs automatically in all networks with proof of work, so that the validation of the blocks occurs in the given time frames and not much faster.
However, since the hashrate went up so much, the difficulty reached the total that the difficulty bomb set, determined by the sum of the difficulty of all the blocks that make up the chain. This is a value that can be checked on Etherscan, as seen in the image below:
When the difficulty bomb was activated, the execution layer (equivalent to Ethereum 1.0) “interpreted” that the proof of work was no longer necessary. However, since there is no Beacon Chain and the Bellatrix fork is not active, there was no consensus layer to “take over” and continue the task of validating new blocks.
How was it fixed?
To end the problem, Ropsten miners were asked to set the TDD to a much higher valueas detailed by developer Marius Van Der Wijden in a publication On twitter. Later, when Bellatrix is activated, you will need to reset it to the actual values. Thus, the hashrate was restored and Ropsten can continue to function normally.
The episode could be interpreted as a malicious attack on the network or as a timely warning. Being a testnet, the cost to the miner that “caused the problem” was not too high (“a few hundred dollars per day”, according to @functi0nZer0), and with that it could have highlighted a possible flaw for the future. the upcoming tests of the Ethereum 2.0 merger.
It should be noted, however, that something like this could not happen in Ethereum 2.0. This is because the Beacon Chain of that network not only already exists, but has been active since 2020 and has more than 300,000 validators awaiting the final merger.
Ropsten, weeks from the merger
As CriptoNoticias reported, Ropsten will have its fusion next June 8. This means that the testnet will stop working with proof of work and switch to using proof of stake to create new blocks.
This is a key test for the subsequent merger on the Ethereum mainnet, which does not yet have a definite date, but is expected this year, according to the developers. Meanwhile, the developers continue to work on improving the stability of the Ethereum 2.0 mainnet, which had its first involuntary fork a few hours ago, although without major consequences.