failover.sh uptime check is questionable

[cmayer68]

the NOTFAILED setting in failover.sh is set to 3600. what is the reason for this setting? it caused
forced breaks with good db replication at customer site when doing the following test:

1. final replicate a->b
2. start watchdog on b
3. stop appserver on a
4. watchdog triggers failover.

if interval from 1 to 3 is > 3600 seconds, no replication break.
otherwise replication break.