RE: nomenclature standardization: where best to put oraInventory?

  • From: "Mark W. Farnham" <mwf@xxxxxxxx>
  • To: <oracle-l@xxxxxxxxxxxxx>
  • Date: Wed, 11 Aug 2004 13:16:51 -0400

I don't think Tanel meant he pointed /etc/oraInst.loc at a non-standard
location, but rather he just editted it to point at the one in ORACLE_BASE
that was the current subject. I could be wrong (easily and often).

So it seems this is indeed a fragile structure. Either it is as officially
supported, one oraInventory in a particular ORACLE_BASE for all ORACLE_BASEs
on a given node, or else you have to tinker.

Jurijs -- what sort of problems did you have? Was it just that oraInventory
was not under ORACLE_BASE, and some patch set made that assumption instead
of consulting /etc/oraInst.loc? I'm thinking that is just a flat-out bug and
someone needs a spanking. However, you might be implying something more
subtle or insidious regarding oui and/or jre complexities and it flew over
my head. Please elaborate if that is the case. If there are fundamental
problems, then I think it means that there is a collision of hierarchical
requirements. I'll try to make an example, but consider this theoretical,
not factual. Say one release of oui needed one release of X services and
another needed a different release, but neither worked with the other and
you could only have one X up on the node at a time. Then the two
ORACLE_BASEs would be incompatible. So for any component that is unique on
the system, if two releases of Oracle require different versions of the
component, then two ORACLE_BASEs could be irreconcilably incompatible.

No one has chimed in on the CRS Home issue yet. Possibly I'm premature at
that, but it seems to me a good solution with well-reasoned nomenclature
needs to be worked out before we can even approach grid free play
substitution and reallocation of nodes. Once upon a time there was a guy at
Oracle who realized that figuring out the lay of the land often took longer
than solving actual problems, so he offered up a decent standard. He cared
about the nomenclature making sense, being flexible enough to actually be
used, and, strangely enough in our relational world, that hierarchies nested
and bound properly.

In this context, I consider oraInventory a current annoyance, and thank you
gentlemen for decent workaround strategies.

But I see CRS Home as a huge debilitating issue that will prevent maturation
of Oracle's grid architecture and complicate actual deployment unless
concensus is reached early in the game. Given that a node might be a
candidate to be deployed to two different services managed by two different
Oracle DBA teams, I find the requirement that CRS Home is under some
specific ORACLE_BASE to be constraining. Clearly for now I'll just have to
live with it. For now I'll probably keep to one ORACLE_BASE per node and
suggest more nodes if group access considerations would otherwise require
more ORACLE_BASEs.

If all y'all have a good idea on this, please chime in.

mwf

-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx
[mailto:oracle-l-bounce@xxxxxxxxxxxxx]On Behalf Of J.Velikanovs@xxxxxxxx
Sent: Wednesday, August 11, 2004 11:57 AM
To: oracle-l@xxxxxxxxxxxxx
Subject: Re: nomenclature standardization: where best to put
oraInventory?


Tanel,
I had the same practice before discovered problem with non-default naming
policy for oraInventory, oui & jre directories.
I don?t remember, but some patching staff doesn?t work on non-default
directories.
After that I started to play with mv command before any installation.

Jurijs
+371 9268222 (+2 GMT)
============================================
Thank you for teaching me.
http://otn.oracle.com/ocm/jvelikanovs.html






Tanel Põder <tanel.poder.003@xxxxxxx>
Sent by: oracle-l-bounce@xxxxxxxxxxxxx
11.08.2004 15:52
Please respond to oracle-l

        To:     <oracle-l@xxxxxxxxxxxxx>
        cc:
        Subject:        Re: nomenclature standardization: where best to
put oraInventory?


Hi!

I've also used this approach of having several inventories per server, but
I've not moved the oraInventory directories around - instead just modified
oraInst.loc file under /etc or /var/opt to state which inventory should
the
installer use.

Tanel.

----- Original Message -----
From: <J.Velikanovs@xxxxxxxx>
To: <oracle-l@xxxxxxxxxxxxx>
Sent: Tuesday, August 10, 2004 11:35 PM
Subject: Re: nomenclature standardization: where best to put oraInventory?


> Mark,
> .
> You have asked question on conceptual level.
> I haven?t answer to the question.


----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to:  oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------


----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to:  oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------


----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to:  oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------

Other related posts: