-2

I created a cluster with two nodes using corosync, pacemaker, heartbeat. This is a HA cluster that host asterisk. Right now I have two resources configured, an asterisk server and a virtual ip address.

cluster conf

All is going smoothly but I have a deficiency. By default when node1 is up and running, the asterisk service in node2 is down. I want to keep both asterisk alive(node1, node2) and control which one is rendering the service with a virtual ip address. I need to do this because every millisecond is crucial, and I don't want to lose time starting asterisk in node2 when node1 fails

Jason Aller
  • 3,541
  • 28
  • 38
  • 38
user1999854
  • 1
  • Unfortunately, Stack Overflow is for *programming* questions. Yours might be suitable for [sf] or [unix.se]; but check their help pages (too!) before posting there, and make sure you're not posting a duplicate question. – tripleee Jul 24 '18 at 15:57

1 Answers1

0

1) attach ip to both nodes

2) start asterisk on both nodes with bind on that ip

3) setup router which send packets to first or second asterisk depend of state.

But that is usless, becuase sip protocol will not work in "milliseconds" like you want and phones will re-register in minutes.

You should consider memory-mapping technics like vmware ESX provide if you need <1second(if you need milliseconds - you have throw in development something like $1000000 and do developemnt yourself).

arheops
  • 15,544
  • 1
  • 21
  • 27