Hi Sourav,
As mentioned by others, doing application failover at DNS level is not a
good idea, whether it is manual or automated (many things like caching is
out of control, and don't forget that in case of power cut in your data
center, many other components will be affected, like switches, DNS,
servers...).
There are many solutions at Oracle level (see some examples:
https://www.oracle.com/technetwork/database/application-development/jdbc/learnmore/trn4065-5180521.pdf)
based on services (that you start depending on the role from startup
trigger, or better: Grid infrastructure or Oracle Restart).
When not possible at client level, Connection Manager can help (note that
application continuity and traffic director need ADG or RAC licensed)
Automatic Failover (FSFO) is a very good solution when you can have the
standby in SYNC, which means low latency between the two sites and/or not a
crazy commit rate. Surely it may do more failovers than what you would done
manually. But that also brings a predictable RTO in minutes. In 19c you can
run FSFO in 'observe only mode' for a while so that you see if you are ok
with it in your context. Reinstate works very well when the databases are
in flashback on. Do not read myths about this - just test it (no restart
required). Of course, there is an overhead but very limited: not all block
changes are logged (only some points to flashback to, and archivelog is
still there to rollforward to precise point-in-time in between those
points) and it is not based on commit but on dbwriter writes (which means
that users do not wait on it).
Most of the automatic failover configurations I've seen had happy DBAs and
happy users as far as the most important recommendation is observed: the
switchover is tested regularly to be sure that all applications follow
automatically. The best for that when the two sites are equivalent (in
capacity and latency) is to change the roles between the two sites, like
every 6 months, when you have a maintenance on the OS or hardware for
example. Then you 1) reduce the outage for that maintenance and 2) you know
that the application follows. Manual or Automatic, you need to trust the
failover and doing frequent switchovers is the best way to keep this level
of trust.
Regards,
Franck.
On Mon, Aug 19, 2019 at 10:41 AM Sourav Biswas <biswas.sourav@xxxxxxxxxxx>
wrote:
Hi Everybody,
This is a client requirement related to Oracle Dataguard.
DG Broker is configured to perform "Failover". Which works fine and fails
over to Physical Standby when the conditions are met. Now, the client
wants, that once DG Broker performs "Failover" its should also notify the
Application DNS Server to redirect its traffic to new Primary Database.
This is part of automation workflow, where client wants to ensure there is
no manual intervention with Database Failover as well as Application switch
to new Primary Database.
Please suggest.
Thanks and Regards,
Sourav Biswas
+91-9650017306
