Comment 3 for bug 1653545

Revision history for this message
Sergei Glushchenko (sergei.glushchenko) wrote :

The cause for this failure is that MySQL skipping redo logging for some DDL operations (https://dev.mysql.com/doc/refman/5.7/en/sorted-index-builds.html). Currently, if you run such DDL during the backup, xtrabackup will abort to prevent creation of corrupted backup.

We are working towards making xtrabackup to retry copying of altered tables or making it block DDLs on server for the duration of the backup.

Current workaround is to retry the backup manually and try to avoid DDLs during the backup.