Implement Migration for Unattested Epochs in Attesting History Database #8121
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
21 tasks
Comment on lines
141
to
162
| // Migrates to a safer format where unattested epochs (source: 0, signing root: 0x0) are marked by | ||
| // FAR_FUTURE_EPOCH as a better default, allowing us to perform better slashing protection. | ||
| func markUnattestedEpochsCorrectly(hd EncHistoryData) (EncHistoryData, error) { | ||
| if err := hd.assertSize(); err != nil { | ||
| return nil, err | ||
| } | ||
| // Navigate the history data by HISTORY_SIZE chunks. | ||
| for i := latestEpochWrittenSize; i < len(hd); i += historySize { | ||
| sourceEpoch := bytesutil.FromBytes8(hd[i : i+sourceSize]) | ||
| signingRoot := make([]byte, 32) | ||
| copy(signingRoot, hd[i+sourceSize:i+historySize]) | ||
| // If source is 0 and signing root is 0x0, we replace that source with FAR_FUTURE_EPOCH | ||
| // which means that epoch in the slice is not yet attested for. | ||
| if sourceEpoch == 0 && bytes.Equal(signingRoot, params.BeaconConfig().ZeroHash[:]) { | ||
| copy( | ||
| hd[i:i+sourceSize], | ||
| bytesutil.Uint64ToBytesLittleEndian(params.BeaconConfig().FarFutureEpoch), | ||
| ) | ||
| } | ||
| } | ||
| return hd, nil | ||
| } |
There was a problem hiding this comment.
Want to put this in the migration file? At some point, we would delete an old migration and having the code co-located would make it easier, i think.
prestonvanloon
approved these changes
Dec 15, 2020
rauljordan
added a commit
that referenced
this pull request
Jan 22, 2021
* Change LowestSignedProposal to Also Return a Boolean for Slashing Protection (#8020) * amend to use bools * ineff assign * comment * Update `LowestSignedTargetEpoch` to include exists (#8004) * Replace highest with lowerest * Update validator/db/kv/attestation_history_v2.go Co-authored-by: Raul Jordan <raul@prysmaticlabs.com> * Update validator/db/kv/attestation_history_v2.go Co-authored-by: Raul Jordan <raul@prysmaticlabs.com> * Invert equality for saveLowestSourceTargetToDB * Add eip checks to ensure epochs cant be lower than db ones * Should be less than equal to * Check if epoch exists in DB getters * Revert run time checks Co-authored-by: Raul Jordan <raul@prysmaticlabs.com> Co-authored-by: prylabs-bulldozer[bot] <58059840+prylabs-bulldozer[bot]@users.noreply.github.com> * Export Attesting History for Slashing Interchange Standard (#8027) * added in att history checks * logic for export * export return nil * test for export atts * round trip passes first try! * rem println * fix up tests * pass test * Validate Proposers Are Not Slashable With Regard to Data Within Slasher Interchange JSON (#8031) * filter slashable blocks and atts in same json stub * add filter blocks func * add test for filtering out the bad public keys * Export Slashing Protection History Via CLI (#8040) * include cli entrypoint for history exports * builds properly * test to confirm we export the data as expected * abstract helpers properly * full test suite * gaz * better errors * marshal ident * Add the additional eip-3076 attestation checks (#7966) * Replace highest with lowerest * Update validator/db/kv/attestation_history_v2.go Co-authored-by: Raul Jordan <raul@prysmaticlabs.com> * Update validator/db/kv/attestation_history_v2.go Co-authored-by: Raul Jordan <raul@prysmaticlabs.com> * Invert equality for saveLowestSourceTargetToDB * Add eip checks to ensure epochs cant be lower than db ones * Should be less than equal to * Check if epoch exists in DB getters Co-authored-by: Raul Jordan <raul@prysmaticlabs.com> * Add EIP-3076 Invariants for Proposer Slashing Protection (#8067) * add invariant for proposer protection * write different test cases * pass tests * Add EIP-3076 Interchange JSON CLI command to validator (#7880) * Import JSON CLI * CLI impotr * f * Begin adding new commands in slashing protection * Move testing helpers to separate packae * Add command for importing slashing protection JSONs * fix import cycle * fix test * Undo cleaning changes * Improvements * Add better prompts * Fix prompt * Fix * Fix * Fix * Fix conflict * Fix * Fixes * Fixes * Fix exported func * test func * Fixes * fix test * simplify import and standardize with export * add round trip test * true integration test works * fix up comments * logrus * better error * fix build * build fix * Update validator/slashing-protection/cli_export.go Co-authored-by: Preston Van Loon <preston@prysmaticlabs.com> * Update validator/slashing-protection/cli_import.go Co-authored-by: Preston Van Loon <preston@prysmaticlabs.com> * fmt Co-authored-by: Raul Jordan <raul@prysmaticlabs.com> Co-authored-by: Preston Van Loon <preston@prysmaticlabs.com> * Filter Slashable Attester Public Keys in Slashing Interchange Import (#8051) * filter slashable attesters from the same JSON * builds * fix up initially broken test * circular dep * import fix * giz * added in attesting history package * add test for filter slashable attester keys * pass tests * Save Slashable Keys to Disk in the Validator Client (#8082) * begin db funcs * add in test and bucket * gaz * rem changes to import * ineff assign * add godoc * Properly Handle Duplicate Public Key Entries in Slashing Interchange Imports (#8089) * Prevent Blacklisted Public Keys from Slashing Protection Imports from Having Duties at Runtime (#8084) * tests on update duties * ensure the slashable public keys are filtered out from update duties via test * begin test * attempt test * rename for better context * pass tests * deep source * ensure tests pass * Check for Signing Root Mismatch When Submitting Proposals and Importing Proposals in Slashing Interchange (#8085) * flexible signing root * add test * add tests * fix test * Preston's comments * res tests * ensure we consider the case for minimum proposals * pass test * tests passing * rem unused code * Set Empty Epochs in Between Attestations as FAR_FUTURE_EPOCH in Attesting History (#8113) * set target data * all tests passing * ineff assign * signing root * Add Slashing Interchange, EIP-3076, Spec Tests to Prysm (#7858) * Add interchange test framework * add checks for attestations * Import genesis root if necessary * flexible signing root * add test * Sync * fix up test build * only 3 failing tests now * two failing * attempting to debug problems in conformity tests * include latest changes * protect test in validator/client passing * pass tests * imports * spec tests passing with bazel * gh archive link to spectests using tar.gz suffix * rev * rev more comment changes * fix sha * godoc * add back save Co-authored-by: Raul Jordan <raul@prysmaticlabs.com> * Implement Migration for Unattested Epochs in Attesting History Database (#8121) * migrate attesting history backbone done * begin migration logic * implement migration logic * migration test * add test * migration logic * bazel * migration to its own file * Handle empty blocks and attestations in interchange json and sort interchange json by public key (#8132) * Handle empty blocks and attestations in interchange json * add test * sort json * easier empty arrays * pass test Co-authored-by: Raul Jordan <raul@prysmaticlabs.com> * builds * more tests finally build * Align Slashing Interchange With Optimized Slashing Protection (#8268) * attestation history should account for multiple targets per source * attempt at some fixes * attempt some test fixes * experimenting with sorting * only one more failing test * tests now pass * slash protect tests passing * only few tests now failing * only spec tests failing now * spec tests passing * all tests passing * helper function for verifying double votes * use helper * gaz * deep source * tests fixed * expect specific number of times for domain data calls * final comments * Batch Save Imported EIP-3076 Attestations (#8304) * optimize save * test added * add test for sad path Co-authored-by: prylabs-bulldozer[bot] <58059840+prylabs-bulldozer[bot]@users.noreply.github.com> * revert bad find replace * add comment to db func Co-authored-by: terence tsao <terence@prysmaticlabs.com> Co-authored-by: prylabs-bulldozer[bot] <58059840+prylabs-bulldozer[bot]@users.noreply.github.com> Co-authored-by: Ivan Martinez <ivanthegreatdev@gmail.com> Co-authored-by: Preston Van Loon <preston@prysmaticlabs.com> Co-authored-by: Shay Zluf <thezluf@gmail.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What type of PR is this?
What does this PR do? Why is it needed?
When we mark epochs as attested for in the validator's attesting history, and we have a gap in epochs, for example:
Then, every single epoch in between is just empty, 0 padded data. This makes it hard to differentiate whether a validator actually attested source epoch 0 and signing root 0x0 or if it is simply unattested for.
To resolve this, we verified there were no attestations with source 0 and signing root 0x0 in mainnet, then we implement a migration logic that marks all those unattested epochs as FAR_FUTURE_EPOCH in this PR.
Which issues(s) does this PR fix?
Part of #7813