hive-engine nodes are replaying

The hardfork didn't go exactly as planned. That's because along with the planned fork to 24 there was a mini fork on 23 about an hour or so before the 24 fork.

As the hardfork delay occurred there was a mix of 23s and 24s running at the same time. There appears to be a block that 24 nodes and 23 nodes didn't agree on and thus a fork happened as a block was rejected by some of the nodes. I haven't seen anyone clearly state why the block was rejected.

It seems that at a later point in what hive-engine recorded there were transactions from the fork that were later merged back into the chain a second time. This is obviously not supposed to happen and the node stopped.

To solve this we have to replay the core hive-engine node and replay the account history node. This should resolve all the problems and ignore the unintended forking.

The good news is that the core node had a snapshot taken 6 days ago. So, that one shouldn't be too bad. We're estimating 22 hours for the replay there.

We're working to figure out the account history side of things. We don't have a working snapshot. We're currently attempting to trim blocks back from the block data we have and we'll start replaying from a few hours before the mini fork. We think this should work, but we'll find out later today.

The hive-engine.com and tribaldex.com sites are in maintenance mode. Deposits/withdrawals that are put on chain now will be stored for the moment, and then sent out at a later time when the main node catches up.

To recap the main node should be back in 24 hours. If the account history block trimming works that should be done sooner. If it doesn't work we'll update as we know more.

Please hold off on any support tickets requests until a day or two have gone by. At that point we'll know whether your transactions were simply delayed or there are problems.