Just been playing with Percona’sXtraBackup and the initial results are very good. I have been using it to populate new slaves from a master with no real downtime.
I found a couple of bugs in the innobackupex script which have been reported. I needed to change line 1760 to
my $wildcard = ‘*.{frm,MYD,MYI,MRG,TRG,TRN,opt,par,ARZ,ARM}’;
Update : This bug has been fixed in the recent 0.9 version
As we use both partitioned and archive tables (it appears the bug also exists in Innodb’s version as well)
The slaves were set up with clean installs of MySQL 5.1.35 (centos5 64bit) and the my.cnf was copied over changing the server-id.
On the host I ran
innobackupex-1.5.1 –stream=tar /tmp/ –slave-info -password=????? | ssh USERNAME@HOSTNAME “tar xfi – -C /var/lib/mysql”
When this had completed on the slave I ran
innobackupex-1.5.1 –apply-log –use-memory=2G /var/lib/mysql
MySQL was then started on the slave and the slave process started using the information from the xtrabackup_slave_info file created on the slave.
CHANGE MASTER TO MASTER_HOST=’????????’, MASTER_USER=’?????’, MASTER_PASSWORD=’??????’, MASTER_LOG_FILE=’mysql-bin.005959′, MASTER_LOG_POS=64860928
SLAVE START
Results are looking good so far