RE: Non-CDB officially deprecated ...

  • From: "Mark W. Farnham" <mwf@xxxxxxxx>
  • To: <iggy_fernandez@xxxxxxxxxxx>, <kellyn.potvin@xxxxxxxxx>, <curtisbl@xxxxxxxxx>, <fuzzy.graybeard@xxxxxxxxx>
  • Date: Thu, 22 Jan 2015 16:12:23 -0500

Of course they (the EM12c team) are doing the right thing.

 

But even though the support matrix is correct, if you're (Oracle) going to
publish a laundry list of unsupported features as reason for delaying going
CDB in conjunction with a deprecation notice (bound to get more attention
than it deserves) then you'd better have EM12c repository hosting on the
list until it is certified (a higher standard than mostly works).

 

None of which, I think, is in the EM12c teams hands, but I think hand
slapping is not in order.

 

Applause will be in order when the kicking the tires is complete.

 

 

From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx]
On Behalf Of Iggy Fernandez
Sent: Thursday, January 22, 2015 3:56 PM
To: mwf@xxxxxxxx; kellyn.potvin@xxxxxxxxx; curtisbl@xxxxxxxxx;
fuzzy.graybeard@xxxxxxxxx
Cc: oracle-l@xxxxxxxxxxxxx
Subject: RE: Non-CDB officially deprecated ...

 

Of course, I meant to say that the EM12c team is doing the right thing.



  _____  

From: iggy_fernandez@xxxxxxxxxxx
To: mwf@xxxxxxxx; kellyn.potvin@xxxxxxxxx; curtisbl@xxxxxxxxx;
fuzzy.graybeard@xxxxxxxxx
CC: oracle-l@xxxxxxxxxxxxx
Subject: RE: Non-CDB officially deprecated ...
Date: Thu, 22 Jan 2015 12:39:00 -0800

I think that the 12c team is doing the right thing in testing before
certifying such a complex product. But FWIW, I would like to quote Bryn
Llewellyn (pronounced Lou-ellen), the most cerebral Oracle employee that I
know, who wrote the white paper on the multi-tenant option
()http://www.oracle.com/technetwork/database/multitenant-wp-12c-1949736.pdf)
.

 

"From the point of view of the client connecting via Oracle Net, the PDB is
the database. A PDB is fully compatible with a non-CDB. We shall refer to
this from now on as the PDB/non-CDB compatibility guarantee. In other words,
the installation scheme for an application backend that ran without error
against a non-CDB will run, with no change, and without error, in a PDB and
will produce the same result. And the run-time behavior of client code that
connects to the PDB holding the application backend will be identical to
that of client code that connected to the non-CDB holding this application
backend. It is intended that a PDB be used to hold a single application
backend. In this way, the PDB provides a direct, declarative means to
contain an application backend so that the Oracle system knows explicitly
which artifacts belong to which application backend. In contrast, when
schema-based consolidation is used within a non-CDB, the Oracle system has
no information about where the various artifacts belong. We shall see that
each foreground process sees, at a moment, just a single PDB; indeed, in the
simplest usage model, a database user, defined within a PDB, is able to
create only sessions that see only that PDB - and only, therefore, the
artifacts that belong to a single application backend.

 

Recognition of the PDB/non-CDB compatibility guarantee allows very many
questions to be answered by assertion: if the answer were not "yes", then
the principle would not have been honored. Tests, conducted by Oracle
Corporation engineers throughout the whole of the 12.1 development cycle
have proved that client code connecting with Oracle Net cannot distinguish
between a PDB and a non-CDB."


Iggy

 

P.S. The underlining of the words "by assertion" was done by me; it was not
in Bryn's paper.

 

  _____  

From: mwf@xxxxxxxx
To: kellyn.potvin@xxxxxxxxx; curtisbl@xxxxxxxxx; fuzzy.graybeard@xxxxxxxxx
CC: oracle-l@xxxxxxxxxxxxx
Subject: RE: Non-CDB officially deprecated ...
Date: Thu, 22 Jan 2015 15:23:58 -0500

So. when you successfully finish your tests then it will be okay that
"Hosting an EM12c Repository" is omitted from the list under the headline

 

"Which features are not supported at the moment?"

 

and until then it would make sense to list it there so folks don't get
themselves into trouble with an unsupported EM12c configuration.

 

Right?

 

And YES, deprecation is a long term warning of future direction, not a panic
button!

 

mwf

 

From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx]
On Behalf Of Kellyn Pot'vin
Sent: Thursday, January 22, 2015 1:20 PM
To: curtisbl@xxxxxxxxx; fuzzy.graybeard@xxxxxxxxx
Cc: oracle-l@xxxxxxxxxxxxx
Subject: Re: Non-CDB officially deprecated ...

 

OK, now Bobby has forced my hand into the conversation and now I have to
slap his... :)

 

1.  Please read the full disclosure-  It simply documents the direction for
the next release.

2.  The final testing is going on for PDBs with EM12c.  There is just the
simple desire to ensure that all tiers of the framework, plugins and 100's
of moving parts are tested thoroughly.

 

Don't make me come over there... :)

 


 


 Kellyn Pot'Vin on about.me
<http://d3mod6n032mdiz.cloudfront.net/thumb2/d/b/a/dbakevlar/dbakevlar-105x7
0.jpg> 

 

Kellyn Pot'Vin-Gorman

about.me/dbakevlar


 

's jus  


~Tombez sept fois, se relever huit!

 

On Thursday, January 22, 2015 11:03 AM, Bobby Curtis <curtisbl@xxxxxxxxx>
wrote:

 

Thank makes no sense, especially with OEM (OMR) not being able to be
supported on a PDB.  Interesting though.

 

 

Bobby Curtis, MBA
curtisbl@xxxxxxxxx
http://about.me/dbasolved



 

On Jan 22, 2015, at 12:39, Hans Forbrich <fuzzy.graybeard@xxxxxxxxx> wrote:

 

I suspected this was going to happen, but was just made aware of the
official statement (thanks Mike and Christian for highlighting the info)

From the 12c "Database Upgrade Guide" at
https://docs.oracle.com/database/121/UPGRD/deprecated.htm#BABDBCJI

"The non-CDB architecture is deprecated in Oracle Database 12c, and may be
desupported and unavailable in a later Oracle Database release. Oracle
recommends use of the CDB architecture."

(There is also a note about features that are not ready in CDB architecture
at the same location.)

I especially like Mike Dietrich's discussion at
https://blogs.oracle.com/UPGRADE/entry/non_cdb_architecture_of_oracle

/Hans

 

 

Other related posts: