Skip to content
This repository was archived by the owner on Jan 22, 2025. It is now read-only.

Fix cluster collapse due to no proper shifted read#7797

Merged
ryoqun merged 2 commits intosolana-labs:masterfrom
ryoqun:fix-bank-hash-mismatch
Jan 15, 2020
Merged

Fix cluster collapse due to no proper shifted read#7797
ryoqun merged 2 commits intosolana-labs:masterfrom
ryoqun:fix-bank-hash-mismatch

Conversation

@ryoqun
Copy link
Copy Markdown
Contributor

@ryoqun ryoqun commented Jan 14, 2020

@sakridge
TBD; My stamina is exhausted; there is moderate number of issues/PRs to submit and tests to write; Those are needed to get here... It was very long journey... ;)

Introduced by months ago: https://github.com/solana-labs/solana/pull/5573/files#diff-2099c5256db4eb5975c8834af38f6456R782.

It seems TdS DR6 finally managed to expose this bug.

The findings at #7736 are now all clear; Every mysterious pieces are put together in place.
And no wonder if this thing happens randomly on several validators with enough stake at the same time, this indeed will lead to mismatched bank hashes across the cluster and the inevitable total consensus failure.

Fixes #7736

@ryoqun ryoqun added the v0.22 label Jan 14, 2020
@ryoqun ryoqun requested a review from sakridge January 14, 2020 20:41
@mvines
Copy link
Copy Markdown
Contributor

mvines commented Jan 14, 2020

nice work @ryoqun!

@ryoqun
Copy link
Copy Markdown
Contributor Author

ryoqun commented Jan 14, 2020

ref: #5573

@ryoqun ryoqun mentioned this pull request Jan 14, 2020
@t-nelson
Copy link
Copy Markdown
Contributor

Good sleuthing!

@sakridge
Copy link
Copy Markdown
Contributor

Great find!

@codecov
Copy link
Copy Markdown

codecov Bot commented Jan 14, 2020

Codecov Report

Merging #7797 into master will decrease coverage by <.1%.
The diff coverage is 100%.

@@           Coverage Diff            @@
##           master   #7797     +/-   ##
========================================
- Coverage    81.9%   81.9%   -0.1%     
========================================
  Files         241     238      -3     
  Lines       51133   51130      -3     
========================================
- Hits        41902   41898      -4     
- Misses       9231    9232      +1

@ryoqun ryoqun merged commit b16c30b into solana-labs:master Jan 15, 2020
mergify Bot pushed a commit that referenced this pull request Jan 15, 2020
* Fix cluster collapse due to no proper shifted read

* Add test for bank hash mismatch

Co-authored-by: sakridge <sakridge@gmail.com>
(cherry picked from commit b16c30b)
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

bank hash mismatch causing dropped votes

4 participants