Critical Description
In case of incremental rebuild/verify we prepare a skip scan using page size constraints first and then use it for rebuild, however when using this skip scan we constraint the number of indexmutations we're going to build/verify also on the page size. In the case where we have multiple index mutations for single dataTableRow, we can return back to the client without scanning all the rows of the skip scan filter, when the client does next, we resume incorrectly from the row after the last of the previous skip scan when infact we hadn't exhausted it fully.
Attachments
Attachments
Issue Links
- links to
