Got a three node cluster.
A backup process takes a full backup weekly and incremmentals daily.
The host choice is from a known members list, application goes into the list in normal order the backup goes in the reverse direction so unlikely to get a backup/application clash. HA proxy selects which node application is connected to.
Now if the full and the ALL the incrementals were taken from the same node recovery works. If the nodes change between the full backup and the incremental the rcovery fails…
Cannot find anything that explains how a backup recovery on an XtraDB cluster should be done… and what extra steps might be needed if the host the full backup was taken from is not avaiable when an incremental is taken.
This supposed to be a HA solution but that also means you have to be able to recover from failure.
The answer must be eluding me… anyone?