Since upgrading to 10.3, I have not been able to get my Production databases backed up using mariabackup. I have 40 servers, 1TB to 30TB, most servers have thousands of identical databases. Here is a snipnet of one failure, however mariabackup has failed on other tables, and I am able to do a select * from the table without a crash.
The backup fails about an hour to 2 hours into the backup.
One thing in common with all the tables reporting corruption either have Data Types of TEXT or LONGTEXT. I also lftp'ed one image to a test and upgraded it to 10.4.13.
 2020-08-07 22:57:50 Database page corruption detected at page 32425, retrying...
 2020-08-07 22:57:50 >> log scanned up to (291691141754836)
 2020-08-07 22:57:50 Error: failed to read page after 10 retries. File ./app_15303/postmeta.ibd seems to be corrupted.
2020-08-07 22:57:50 0 [Note] InnoDB: Page dump in ascii and hex (8192 bytes):
InnoDB: End of page dump
2020-08-07 22:57:50 0 [Note] InnoDB: Compressed page type (11); stored checksum in field1 4040847758; calculated checksums for field1: crc32 2237276139, innodb 4040847758, none 3735928559; page LSN 129654058625760; page number (if stored to page already) 32425; space id (if stored to page already) 587011
The command I used:
/usr/bin/mariabackup --defaults-file=/etc/my.cnf --socket=/var/lib/mysql/mysql.sock --use-memory=64G --user=USER --password=PASSWORD --parallel=12 --no-lock --kill-long-queries-timeout=60 --tmpdir=/database/tmp --stream=xbstream --backup --target-dir /database/production/ | nc XXX.XXX.XXX.XXXX 9999