Fix incremental backup interaction with XLOG_DBASE_CREATE_FILE_COPY.
authorRobert Haas <[email protected]>
Mon, 4 Mar 2024 18:33:12 +0000 (13:33 -0500)
committerRobert Haas <[email protected]>
Mon, 4 Mar 2024 18:33:28 +0000 (13:33 -0500)
commitd75c4027b6f260f2045b162017567aeeb909b056
tree35488e90b82274595b76e1484b2b582d7b779cba
parentcb6945dc8061d03e6ec670a6856228f12e94264c
Fix incremental backup interaction with XLOG_DBASE_CREATE_FILE_COPY.

After XLOG_DBASE_CREATE_FILE_COPY, a correct incremental backup needs
to copy in full everything with the database and tablespace OID
mentioned in that record; but that record doesn't specifically mention
the blocks, or even the relfilenumbers, of the affected relations.
As a result, we were failing to copy data that we should have copied.

To fix, enter the DB OID and tablespace OID into the block reference
table with relfilenumber 0 and limit block 0; and treat that as a
limit block of 0 for every relfilenumber whose DB OID and tablespace
OID match.

Also, add a test case.

Patch by me, reviewed by Noah Misch.

Discussion: http://postgr.es/m/CA+Tgmob0xa=ByvGLMdAgkUZyVQE=r4nyYZ_VEa40FCfEDFnTKA@mail.gmail.com
src/backend/backup/basebackup_incremental.c
src/backend/postmaster/walsummarizer.c
src/bin/pg_combinebackup/meson.build
src/bin/pg_combinebackup/t/006_db_file_copy.pl [new file with mode: 0644]