I set the VRA 7 distributed form, I 2 appliance VRA and 2 IaaS Server. These servers are configured in a load balancer, nuvem.s4c.local (points to both VRA VA) and web.s4c.local / manager.s4c.local (points to both IaaS Server).
S4C-VRA7A
S4C-VRA7B
S4C-IAAS7A
S4C-IAAS7B
Go up the environment, set up a few things, I created blueprints and did deploy virtual machines. To validate this HA environment I turned off all machines "B" (S4C-VRA7B and S4C-IAAS7B), this time some virtual machines disappeared from the inventory of the VRA, but remain in the vCenter IP and remained registered in the network profile.
My question is: it is a database replication issue between the VRA appliances or between IaaS Server?
I know that replication between the VRA appliance are asynchronous by default, but can be configured to function as synchronous is recommended? Will I lose performance?
How do I know if replication is happening after stopping a server?
Another detail is when a VA VRA this offline, the other VRA VA is promoted to master, however, when the old master is online both VRA VA are as masters (each in its console), which procedure so that both know who is the master? The documentation is not clear about it
Thank you very much.