That would explain it. I have no doubt it is documented some where, but I
havent found it yet.
On Thu, Dec 10, 2015 at 8:55 AM, TJ Kiernan <tkiernan@xxxxxxxxxxx> wrote:
This is the behavior I’ve seen (11.2 on Windows) as well and assumed it’s
normal. If you intentionally relocate the database (srvctl relocate
database -d <dbname> -n <nodename>), instance_2 will start before
instance_1 stops, but in a failure, instance_1 starts wherever it can.
Unless the 10 day rule is enforced via instance name (_1 vs _2), I don’t
think it’d be a problem. Maybe running instance_2 on node1 looks
confusing, but there shouldn’t be any difference in operation.
Thanks,
T. J.
*From:* oracle-l-bounce@xxxxxxxxxxxxx [mailto:
oracle-l-bounce@xxxxxxxxxxxxx] *On Behalf Of *Andrew Kerber
*Sent:* Wednesday, December 09, 2015 7:29 PM
*To:* ORACLE-L
*Subject:* rac one node naming
Some time ago we had a discussion of rac one node instance naming, and as
I recall the consensus was that the instance name was always the same
(normally _1).
So, I did some experimentation, and I need to understand what is going on
and how to prevent it.
I put together a 2 node rac one node installation in VMware workstation
12, oracle 12c/rhel 6.5.
The running instance is rontest_1. Service name rontest., nodes ractest1
and ractest2, instance running on ractest1
I shut down ractest1, and the rontest_1 instance is now running on
ractest2.
Now, I manually relocate the instance, using the command:
srvctl relocate db -d rontest -n ractest1.
I look at node 1, and now rontest_2 is running on ractest1.
So, what did I do wrong, and how do I prevent it from happening again? Or
is this the expected behavior.
--
Andrew W. Kerber
'If at first you dont succeed, dont take up skydiving.'