Oracle Fail Safe

Can somebody explain how failsafe works from a networking point of view.

Does the primary publish a virtual IP address; the primary and secondary share this IP address and  and MAC address and following fail-over, the secondary just answers the same requests as the primary.

Or do they share an IP address but have different MAC addresses, so a re-ARP is required by the client is required, or is action required by the client to re-initiate a connection to the secondary server after fail-over (i.e. a specially written client is required?

Thanks,
Stevod
Stevod2Asked:
Who is Participating?
 
tomcatkevConnect With a Mentor Commented:
The cluster configuration can have the same Virtual IP for whatever is the active node following failover, but doesn't necessarily have the same MAC address.  This is really a matter of how the Windows cluster is configured, and I don't think you want multiple active nodes with the same MAC address, and if you are doing some Active / Active clustering (i.e. 1 database on each nodes, with failover resulting in 2 DB running on the remaining active node), then you really can't change MAC addresses, but you can set up seperate VIPs for the 2 databases, and have one node that accepts connections for both VIPs.

It is possible to code TNSNAMES.ora to achieve Transparent Application Failover, this may be something you want to look into...  http://wiki.oracle.com/page/Transparent+Application+Failover+(TAF)
0
All Courses

From novice to tech pro — start learning today.