xtrabackup 2.3 behavior changes
Affects | Status | Importance | Assigned to | Milestone | ||
---|---|---|---|---|---|---|
Percona XtraBackup moved to https://jira.percona.com/projects/PXB | Status tracked in 2.4 | |||||
2.3 |
Fix Released
|
High
|
Hrvoje Matijakovic | |||
2.4 |
Fix Released
|
High
|
Hrvoje Matijakovic |
Bug Description
I upgraded to xtrabackup 2.3 and my automated scripts are throwing errors in the new version. I believe it's because innobackupex has been merged into xtrabackup, but is there a way to re-create the old behavior? My scripts run xtrabackup directly without going through innobackupex such as:
xtrabackup --backup --datadir=
My issues are:
1) xtrabackup now requires a username/password to run. It previously ran backups without needing this. Is there a reason a username/password is now needed or is there a benefit to supplying it? The documentation doesn't mention a username/password at https:/
2) I use xtrabackup to backup specific tables and then restore them by following the instructions at https:/
Changed in percona-xtrabackup: | |
status: | New → Triaged |
Hi Ryan,
Thank you for detailed report. Yes, innobackupex has been merged into xtrabackup with the intention to get rid of innobackupex and make xtrabackup to do all the things that innobackupex previously did. It means that xtrabackup does now copy all the files (InnoDB, MyISAM, FRM, etc.), it takes care of binary log coordinates, and so on. It was intentional, but our fault that that we did not document this behavior properly.
As for the case when you need to copy only InnoDB tables, we might consider implementing such mode. Note, that we never recommended to xtrabackup directly.