New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Update proof parameters file to most recent version. #1595
Conversation
Codecov Report
@@ Coverage Diff @@
## main #1595 +/- ##
=======================================
Coverage 34.49% 34.49%
=======================================
Files 208 208
Lines 23384 23384
=======================================
Hits 8066 8066
Misses 15318 15318 Continue to review full report at Codecov.
|
Good! What about removing |
It failed to sync for me without it. |
I managed to sync without but maybe we can keep it and be conservative. |
It works for you if you remove |
My bad, I forget to recompile the client after my change (didn't saw the |
Yes, it works for me and the file |
@lemmih @elmattic any wrap-up on why it caused an error, where in the system the artifacts were persisted (and grabbed by Forest)? We should understand this in full lest we risk it happening again. And perhaps add some assertions if it's feasible. On a testing side note, once the nightly check machine is back I'll change it to not build Forest but to take the image from the registry, ensuring clean state. We would be able to catch this class of bugs sooner. |
Can confirm it's not working without |
Brief post-mortem: Guillaume and I were able to sync Forest to mainnet but Hubert saw state root mismatches with the exact same code. No matter how many times we cleared the database and built Forest from scratch, the two different outcomes wouldn't go a way. We faultily assumed all our data was stored in It's a significant problem that a parameter mishap leads to such hard-to-debug state root mismatches rather than fail with informative error messages. Skimming our code as well as that of rust-fil-proofs, I don't see any trivial ways of checking whether we have the most up-to-date parameters. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM as a workaround, but this JSON is not really maintained by us, so in my opinion, it should be fetched during e.g. build phase or the relevant repo should be added as a submodule.
Yeah. We can't use the un-edited json file, though. We need to add the entry for "v28-fil-inner-product-v1.srs" (reasons still unknown). |
I'll merge this PR but this is not done. We 100% need something better. |
Summary of changes
Changes introduced in this pull request:
Reference issue to close (if applicable)
Other information and links