You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
the code below should be executed only for reference ballot. the sync is incredibly slow with that loop
also for current epoch we should get this data from tortoise rather then lru. lru is bad here, because if set to 10_000 and in fact there are 11_000 in active set then this cache will be threshed. tortoise on the other hand needs to store this data regardless, and it stores it in a compact form.
The text was updated successfully, but these errors were encountered:
tried to sync node and noticed that it spends a lot of time looping over activeset in eligibility validation. i will open an issue to optimize that code path, but for now we need to increase number of activations stored in cache.
related: #4883
closes: #4883
related: #4841
active set is validated only for reference ballot in current epoch, for all other ballots eligibilities validated against declared slots in the reference ballot. beside making validation more efficient, it will allow to prune active set data in future changes. reference ballots is still loading data from cached datastore, i plan to remove eventually but not in this change.
additionally i rewrote all tests for eligibility validator to get to 100% coverage and make them less dependent on the implementation.
the code below should be executed only for reference ballot. the sync is incredibly slow with that loop
also for current epoch we should get this data from tortoise rather then lru. lru is bad here, because if set to 10_000 and in fact there are 11_000 in active set then this cache will be threshed. tortoise on the other hand needs to store this data regardless, and it stores it in a compact form.
The text was updated successfully, but these errors were encountered: