Partial witness validation: track already observed witnesses #11302
Assignees
Comments
github-merge-queue bot
pushed a commit
that referenced
this issue
Jun 5, 2024
This PR introduces a separate LRU cache to keep track already processes state witnesses. Note that malformed witness (that couldn't be decoded or decompressed) is also considered processed because we want to avoid processing more parts for that chunk. When we have enough witness parts the entry is removed from `parts_cache` and added to `processed_witnesses`. This makes `is_decoded` field in `CacheEntry` redundant since now `parts_cache` only stores not-yet-decoded witnesses, so the field was removed. This PR also includes a bit of refactoring such as getting rid of `CacheEntry` `shard_id` field and removing/renaming metrics. Closes #11302.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Part of #11193
We want to have a separate LRU cache to keep track of already decoded state witnesses. This is needed to prevent attacks on chunk validators by sending the same witness multiple times and making it apply chunks which can be resource-intensive.
Similar to #11111
Also see #11193 (comment)
The text was updated successfully, but these errors were encountered: