HP StorageWorks Storage Mirroring for Linux User's Guide (T2558-96078, February 2008)

11 - 3
4. The failed source’s IP address and subnet mask are added to the target’s identity.
5. Failover sends updates to routers and other machines with IP address mappings allowing clients to seamlessly attach to
the target machine. All user and application requests destined for the source machine are routed to the target machine.
6. If specified, a post-failover script executes on the target. This script is user-defined and optional. This may be used to
start daemons that will be needed after the target assumes the identity of the failed source. Scripts can also send
network messages notifying administrators that failover is complete.
7. Failover is complete.
How Failback Works
1. The network administrator disconnects the failed source machine from the network and repairs it.
2. The network administrator manually initiates failback.
3. If specified, a pre-failback script executes on the target. This script is user-defined and optional. This may be used to stop
daemons on the target that will not be needed or may be in conflict after the target reinstates its own identity. Scripts
can also send network messages notifying administrators that failback is about to occur.
4. The source’s identity is removed from the target.
5. Failback sends updates to routers and other machines with IP address mappings. This mapping is now the original
mapping before failover occurred. At this time, users may see a pause at their workstations while the failback process
completes.
6. If specified, a post-failback script executes on the target. This script is user-defined and optional. This may be used to
start daemons that will be needed after the target reinstates its own identity. Scripts can also send network messages
notifying administrators that failback is complete.
7. Failback is complete and you are given the choice to continue monitoring the source machine.
8. The original source machine can be reattached to the network.
9. If specified, a post-failback script executes on the source. This script is user-defined and optional.
10. Failback is complete.
NOTE: Users who were using the address of the source machine were actually attached to the target machine and
will lose their connection to the server at this point.
WARNING: After the source machine is back online, if you need to restore the data on the original machine by
mirroring data from the target’s copy of the replication set, do not let users logon to the source until the
restore process is complete.
Failed Source
IP Address:
Target Name:
Backup
125.7.82.67 and 125.7.82.43
User and application requests are sent to the target by
locating the IP address