SRX Cluster scenario failover, two ISPs and BGP design problems?
[ Edited ]
We have the attached scenario to solve on the customers site. The part considering the servers is the issue for Linux users, so if someone has advices, please comment.
The customer wants the availability of his server all the time. The server is in fact web server/web application with some mysql db etc., and that part is solving with HA ext on Linux, but still remaining the problem of loosing network connectivity.
In the network part they want two SRX 240 cluster and two different ISPson the central and backup location. For data/control link we will use dark fiber with some 400m length. I have some issue and problems understanding how will it work:
1. I think I can achive the vlan scenario on the picture (vlan 10 for public server where NICs are on active/passive state, and vlan 11 for communication between them) usin Junos 11.4 version, right?
2. Considering cluster: the weak part id dark fiber/control-data link. So if the fiber is down node 0 is active and node 1 is in the disabled state. But what is happening when the node 0 has no power. The node 1 goes also to the disabled state and I have no failover, right? So what is the point of cluster. Is it only that I have reth links and if the node 0 RE is failed for some reason I have backup RE on node 0?
3. The servers must have same public IP and are working in active/disabled mode. Can I have some active/active scenario and internal routing in this case without cluster?
Ok the situation when fiber is down, node 0 is active and thats ok. But the main problem is when node 0 is powered off how to solve that server on the main location goes off and node 1 and server on his location becomes active. If I can do that than cluster isn't the solution.
Is there a way that in any case (data/control link down or power down for node 0) node 1 becomes active?