Gotta love that which “cannot be discussed” 😊
Reminds me of a time long ago and far away (Oracle 7, Oracle Parallel Server) –
we had a system that captured call detail records from a “consolidation server”
and then fanned them out to other systems. One of the “requirements” was that
the data had to get from our server to the other systems within less than an
hour and the records were stored “persistently” on the downstream systems. The
consolidation server stored 5 days of data and it was easy to “replay” the data
from them. So, to improve performance (actually pretty significantly) we had
the database in NOARCHIVE mode. This was a system we sold to various telcos
and we were constantly arguing with their DBAs about why recovery would be
easier and faster to just replay any “lost” data from the consolidation server.
Then, on the flip side, there was another company I worked for where the
“senior” DBA insisted “ARCHIVEMODE is worthless” right up until the time we had
a failure that could have been recovered “if only……”.
Clay Jackson
Database Solutions Sales Engineer
clay.jackson@xxxxxxxxx<mailto:clay.jackson@xxxxxxxxx>
office 949-754-1203 mobile 425-802-9603
[cid:image002.png@01D5985F.0243A3C0]
From: oracle-l-bounce@xxxxxxxxxxxxx <oracle-l-bounce@xxxxxxxxxxxxx> On Behalf
Of Mladen Gogala
Sent: Sunday, November 10, 2019 8:39 AM
To: Mark W. Farnham <mwf@xxxxxxxx>; franck@xxxxxxxxxx
Cc: 'Oracle-L Freelists' <oracle-l@xxxxxxxxxxxxx>
Subject: Re: is there any way i can avoid rebuild physical standby database
after every PROD to UAT database refresh ?
CAUTION: This email originated from outside of the organization. Do not follow
guidance, click links, or open attachments unless you recognize the sender and
know the content is safe.
I clarified the problem in a little private exchange. Falgun has a UAT server
which is protected by the standby. His problem is that whenever he refreshes
the UAT db, he also needs to refresh the standby for the UAT db. My first
question was why was there a standby for a database that is regularly
refreshed? Backup should be protection enough, since the data already exists in
production. However, my impression is that the standby for UAT server cannot be
discussed and is a given. In that case, the answer is a clear "no". Standby
also has to be refreshed. Hopefully, he uses SRDF, HUR or SnapVault to refresh
UAT and standby.
Regards
On 11/10/19 10:28 AM, Mark W. Farnham wrote:
What exactly do you wish to refresh from production?
Can you accomplish your goal with transportable tablespaces and the files that
contain them?
There is probably a difference in approach with the not necessarily equal
efforts of goals for “least human intervention” versus least physical activity
on the servers.
Increasing complexity is usually proportional to increasing failure rates,
although different folks put different metrics on discrete human actions,
number of steps, and underlying complexity.
And at some scale versus any then current variety of physical components,
reasonable time to complete and consumption of resource limited resources may
require additional complexity to complete in time and resource consumption that
is acceptable.
I personally have a distaste for the physical machinery wasting cycles, but
that is something I dismiss depending on the real world situation.
So is avoiding the rebuilds a matter of taste or a real issue of time to
complete?
Clearly stating the goal (including the acceptable service levels and physical
resource capacities) is a useful starting point.
The short answer to your question is “yes,” but a presumed toolset and exactly
what you mean by a physical standby database may limit your options.
Good luck,
mwf
From: oracle-l-bounce@xxxxxxxxxxxxx<mailto:oracle-l-bounce@xxxxxxxxxxxxx>
[mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Franck Pachot
Sent: Saturday, November 09, 2019 4:31 PM
To: Mladen Gogala
Cc: Oracle-L Freelists
Subject: Re: is there any way i can avoid rebuild physical standby database
after every PROD to UAT database refresh ?
Hi Falgun,
Even without resetlogs, you have to rebuild the standby because the primary has
been overwritten.
If you do not open resetlogs (like with a restore without open) then you can
rebuild the standby from the same backups that were used for the primary. But
I'm not sure that it will be better than re-creating the standby from the
primary. Is that a problem if it is automated (duplicate for standby, or
restore from service is easily automated).
Regards,
Franck.
On Sat, Nov 9, 2019 at 1:16 AM Mladen Gogala
<gogala.mladen@xxxxxxxxx<mailto:gogala.mladen@xxxxxxxxx>> wrote:
On 11/8/19 5:05 PM, falgun patel wrote:
I have question related to Dataguard.
I get 2 to 4 times request in a month to refresh Oracle Production Database to
UAT Database.
Production and UAT both databases are on 11g R2 version.
Production Database Setup is 4 Node RAC i.e. 2 Node Primary and 2 Node Physical
Standby.
UAT Database Setup is 4 Node RAC i.e. 2 Node Primary and 2 Node Physical
Standby.
My Question:
Every time if refresh Production database to UAT , and since every time
incarnation gets changed in UAT after refresh from production database.
Due to this my current UAT Physical standby database is of no use as it goes
out of sync.
Everytime After this production to UAT DB refresh i have to rebuild my UAT
physical standby database again and again.
I want to know if there is any way i can avoid this rebuild physical standby
database after every UAT database refresh ?
Please provide me detailed steps.
--
Regards
Falgun Patel
Hi Falgun,
Of course you can avoid the rebuild of physical standby database. Why are
changing the incarnation? Let me guess: you are connecting to catalog. I have
been refreshing databases in the ancient times of 9.2 without changing the
incarnation. You don't connect to catalog and you don't register the newly
minted UAT copy to catalog. Also, use SRDF or whatever tool your array offers.
Regards
--
Mladen Gogala
Database Consultant
Tel: (347) 321-1217
--
Mladen Gogala
Database Consultant
Tel: (347) 321-1217