I understand that in a master failure clients automatically fail over to the slave which becomes the new master. Then we have to manually stop the master & slave & move the slaves data files over to the master then restart the master and slave.
If a master goes down due to hardware failure and we want to bring it back up again on some other machine, take the slave (which is now the master) down, copy its persistent files to another machine and start the master and slave back up again. My question is what happens to the clients during this manual fail over window if they still trying to send messages? My initial test appears client threads are being blocked until the broker available. Is this the expected behavior? if yes... is there any way we could set a timeout/notify client on whats happening..? -- View this message in context: http://www.nabble.com/Recovering-after-failure-in-Master-Slave-configuration-tf3124553.html#a8656729 Sent from the ActiveMQ - User mailing list archive at Nabble.com.
