Handling of compressed tablespaces with compressed page size == server page size broken

[Laurynas Biveinis]

A regression in fix for bug 932623: neither the zip_size nor table compression size is stored in the metadata. Instead, an assumption is made that if page size < UNIV_PAGE_SIZE, then a tablespace is compressed, and if page size == UNIV_PAGE_SIZE, then it's not compressed, e.g. in xtrabackup_apply_delta():

 dst_file = xb_delta_open_matching_space(
   dbname, space_name, info.space_id,
   info.page_size == UNIV_PAGE_SIZE ? 0 : info.page_size,
   dst_path, sizeof(dst_path), &success);

This assumption breaks for compressed tables with page size == UNIV_PAGE_SIZE, and the actual zip_size should be stored in metadata in addition to page_size.

[Laurynas Biveinis (laurynas-biveinis)]

There is no user impact in 2.0.2. It only starts breaking things for other work in progress.

[Shahriyar Rzayev (rzayev-sehriyar)]

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PXB-340