I haven't checked the install docs for a while, but in days of yore the
/etc/hosts method proposed by your infrastructure team used to be required
and/or "strongly recommended" for Oracle Apps deployments.
On Tuesday, 27 October 2015, Alfredo Abate <alfredo.abate@xxxxxxxxx> wrote:
Hello,
In our Oracle E-Business Suite 12.1.3 environments, we have been having
some minor issues when it comes to our DNS servers being rebooted for
Windows patching (both primary and secondary but in a staggered fashion).
During this time the Oracle EBS application servers are momentarily unable
to ping each other and we see this in our Internal Manager logs. They
recover successfully and the whole thing lasts around 2-3 minutes. This is
not an issue now as it happens during a maintenance window when no users
are on the system but that will change in the future when we are a true
24/7 environment. We observe this in both our Production and
Non-Production environments.
Both Oracle EBS and Database are on Linux 6.4. We are testing the timeout
and attempts parameters that are in resolv.conf to see if we can find a
balance for it to fail-over to a secondary DNS server within a reasonable
amount of time. Oracle provides an ebs_server R12 pre-install RPM that we
have installed and sets the values to attempts 5 and timeout 15. This
hasn't seemed to help us but we are continuing to tweak these to see if a
different value helps us.
Our infrastructure team mentioned hard coding the application/database
servers in /etc/hosts so it can resolve the name and won't lose
connectivity when DNS is unavailable. This is not my preferred method but
wanted to see what others thought.
I'm curious to hear if others have experienced this issue and how they
manage dealing with DNS server maintenance/reboots for their Oracle
environments (E-Business and/or Database)?
Thanks,
Alfredo
