We need to set a fault tolerant solution for a VOIP application preferably using Linux clusters.
The application will reside on server A and an identifically configured server B will act as fail over. While we are able to switch over the IP and even ensure that application is started within a minute on failover using HeartBeat & DRBD, there are call drops. The reason is that while data can be dynamically copied using DRBD, the call sessions which I presume are still in memory crashes.
We are able to set up a high availability cluster using Vmware Enterprise editon, but the cost of the solution becomes very prohibitive.
I have been advised that you can set up an Ubuntu Cluster which will be HA solution, using Ubuntu Cloud. I tried the manual, but there was no way we can actually have an instance which can be served across two or more servers, so that even if one machine goes down the Virtual machine stays intact.
What I essentially require is a solution where the virtual machine stays on even if the primary server goes down, the sessions stay stable. Dont suggest heartbeat&DRBD it does'nt work in this case. If its possible using some other open source platform, please advise