Giters

dtube / avalon

Blockchain for social distribution

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Avalon Chain Halt Reason/Solution documentation/discussion

brishtiteveja opened this issue · comments

commented

These are the known reasons for avalon chain halt

  1. Offline: Less than 2/3 online (2/3 offline) with good chain state
    Is it possible for a node to know.. it's in "good chain state" or a bad chain state. If you are not producing for 3 * 15 = 45 sec or last 120 sec, then possibly you are in "bad chain state"

  2. Double production: When a single leader with same leader key tries to mine from multiple device, it causes two different block hash generated in different device and broadcast.. other leader gets confused and some receive from one source.. other receives from other source for the same leader.. thus chain split.

  3. Time limit exceeded: Very slowed down system (for disk usage for example) that slows down a leader's mining action, then the node tries to mine very late at the same time as the backup leader (the previous one). Forking the chain. It happened recently.

Add more reasons.. discuss possible solutions.

Possible solution:
Rollback if hasn't mined for more than 2 minute. Check whether mining becomes normal. Rollback at least 5 times. If it is still stuck, then notify.

commented

If anyone would like a backup of the chain halt before last: /ipfs/QmfBHvKEBhyLBcVygY2Mg8cJwNm6mYBXgn256VpC6j3N7s
Video description of my interpretation of the halt before last: https://d.tube/v/devdeckardcain94/QmasnRj7DkAdMDPHTijAQvmcHHG26NPgNjcpVqFgkRPMNT

This mornings chain halt backup: /ipfs/QmeWpYJrJXdrsWKwJjRTmPBETGN7XKPi6hGcdV6UucC1ex