Giovanni S | Flow (2024-07-25)
Question about block time once block production resumes post-Crescendo - will the onchain time delta between pre- and post-Crescendo blocks reflect the offchain time spanned during migration?
7 Replies
I've created a thread for your message. Please continue any relevant discussion in this thread.
You can rename this thread using
/title <new title>
If this is a technical question that others may benefit from, considering also asking it on Stackoverflow: https://stackoverflow.com/questions/ask?tags=onflow-cadencecc @Jan | Flow @TheOneSock
Or will the block timestamp limitations that currently exist - block(n+1).timestamp - block(n).timestamp <= 10 seconds
Ref: https://cadence-lang.org/docs/1.0/measuring-time#time-on-the-flow-blockchain
I have no idea. Maybe @Jordan S | Flow would know? We can also take a look at past spork boundary logs, I can take a look at that later today.
Also tagging @Peter | Flow and @Sjon-Paul | Flow, they might know.
Unknown User•5mo ago
Message Not Public
Sign In & Join Server To View
When we spork, we set the spork root block timestamp to
time.Now() at the time it is constructed. When it is constructed is early in the spork process, because partner consensus operators need to pull the block down, sign their votes for it, and send the votes to us.
In general the timestamp properties for the first block F of a spork would look like:
- F.timestamp = F.parent.timestamp + k1, k1 is much larger than 10s
- If t_start is the actual time when the network starts after the spork, then F.timestamp = t_start - k2, k2 is much larger than 10s and much larger than k1.
- After the network starts up after the spork, the timestamp field will increase by 10s per block (faster than wall-clock time) until it converges with wallclock time.
(This is all assuming we use the defaults and don't pre-specify the timestamp flag to be the expected startup time as part of the spork process. I'm pretty sure we don't.)Interesting, thanks for the details!
After the network starts up after the spork, the timestamp field will increase by 10s per block (faster than wall-clock time) until it converges with wallclock time.Do we have estimates of how long the HCU will take in meatspace time and how long it will take for block time to converge again?
HCUs would be different from sporks. Blocks are produced continually (just not executed or sealed), so block timestamps would be as accurate as they normally are, w.r.t. block creation time. It would just take longer for their result to be sealed. I think HCUs normally take around 5-10 minutes.