Currently, we have some issues with listing large prefixes. In the past (before metabase) we had optimization that was helping with that but we didn't move it to metabase. This optimization is for non-recursive prefix listing.
Optimization was working in a way that if while listing key contained delimiter (/) then the current listing page was dropped and the cursor for the next page was set to prefix<byte of / + 1>.
Currently, we have some issues with listing large prefixes. In the past (before metabase) we had optimization that was helping with that but we didn't move it to metabase. This optimization is for non-recursive prefix listing.
Optimization was working in a way that if while listing key contained delimiter (
/
) then the current listing page was dropped and the cursor for the next page was set toprefix<byte of / + 1>
.Old implementation can be found here (see
skipPrefix
): https://review.dev.storj.io/c/storj/storj/+/4607/3/storage/cockroachkv/ordered_iterator.goAcceptance Criteria: