I have a small problem. I need to get the data dir of a master server in a cluster to stay in sync with a slave server so when it fails over the data will all be in tact. However, that is not my problem, the problem is with syncing the master's data with the data on the "slave" when the server fails back. DRBD, NBD, ENBD , or any network block device is not an option for me since I can't re-partition my drive in this case (I don't want to mirror all of /var). I've been told to do master-master replication, but I've seen no written documentation on that for MySQL, plus it seems like it could be very prone to corruption. Does anyone have any ideas on what I should do in this case?