Alessandro Ren wrote:

I did that, as soon as I start replication, the relay file gets corrupted and replication stops. []a.

Alessandro Ren wrote:

...

Has someone had problems with the mysql version of Centos 5.2? After 

I upgraded, the replication broke, the relay file got corrupted, I had to downgrade for it to work again.

no such problems here. You could potentially rebase your replica's once you have done the upgrade.

When you 'rebase' there is no relay file on either side. Thats the whole point of doing a rebase from scratch. Essentially, you'd do something like this on the master :

Service mysqld stop; cd /var/lib/mysql/ ; rm -f mysqld-bin.* ; rm master.info; rm relay*.info ; lvcreate -s -n mysql-snap -L 10G /dev/VolGroup00/LogVol00 ; service mysqld start; mkdir /mnt/mysql-snap; mount /dev/VolGroup00/mysql-snap /mnt/mysql-snap; ssh root@replica /sbin/service mysqld stop; rsync --delete -Pvar /mnt/mysql-snap/* root@replica:/var/lib/mysql/ ; ssh root@replica /sbin/service mysqld start

If you still use the old legacy way of specifying the replica param's on the replica machine, things should just work from there. if you specify them on the cli, then login to the replica mysql, and 'change master to master_host...' make sure you set position to '0' since the master would have started writing new binlogs.

Also, dont top post, and trim your posts.