Re: More spfile goodness
- From: Peter Gram <peter.gram@xxxxxxxxxxxx>
- To: oracle-l@xxxxxxxxxxxxx,Peter Gram - Miracle A/S <peter.gram@xxxxxxxxxxxx>
- Date: Tue, 25 May 2004 02:11:57 +0200
Dear Members
I have been following the spfile pros / cons thread and there are many
god arguments for not
to use spfiles.
Regarding the reason for making the spfile binary - Could it be RAC ?
Since a binary file
could reside on a raw device and therefor could be seen by all instances
on a RAC's that does not
have a clustered file system.
I don't have a RAC to test this
Just my 2 cents
/peter
Mogens Nørgaard wrote:
>Ah, a real pro from the old days joins the list!
>
>For those of you who weren't there, Dom was THE replication hero of the
>7.1.6 days. Only one who could make it work.
>
>He was the true 2iC of Cary's System Performance Group. I still recall
>the Teach-The-Trainer (TTT) class for Replication in Dallas (I think -
>or was it Atlanta?) with Dom on a conference call to answer our
>questions, and the heavies from Education present to make sure it was
>all going to be OK.
>
>Welcome, Dom. Good to have you on board.
>
>Mogens
>
>Delmolino, Dominic wrote:
>
>
>
>>My First Post, so newbie apologies in advance :-)
>>
>>We considered, and discarded, the idea of moving to full use of spfiles
>>during our upgrade from 8i to 9i using the following reasoning:
>>
>>Pros
>>
>>- SPFILEs can be modified without the need to log into the server O/S
>>- Online parameter changes could be captured for the next db restart
>>
>>Cons
>>
>>- We version control our init.ora files and the diff capability is
>>useful
>>- The ability to look at the parameter values prior to instance start is
>>
>> useful
>>- You need permission to log into the server O/S in order to modify the=20
>> initialization parameters (security)
>>
>>The Cons were pretty important to us and the Pros seemed to have both
>>security issues (we don't want any DBA to be able to modify the file,
>>we want the added O/S security) and unintential consequence issues=20
>>(Change this parameter for this run -- don't fat-finger SCOPE=3DBOTH
>>by accident). Since the old format was still supported, we stayed with
>>it.
>>(We did take the opportunity to thoroughly scrub the old files and ditch
>>
>>parameters for which we should have been using Oracle's defaults
>>instead).
>>
>>I'm guessing that the reason a binary file was chosen was because the
>>developer used some weird internal object to represent the parameters=20
>>(heck, it's probably a Java-serialized-object :-) and sold his/her
>>management that it was necessary in order to achieve the Pros as listed
>>above :-) ("Don't want those pesky DBA's messing up the file format").
>>
>>Dominic Delmolino
>>
>>
>
>
>----------------------------------------------------------------
>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
>-----------------------------------------------------------------
>
>
>
--
Peter Gram
comp : Miracle A/S
Addr : Kratvej 2, 2760 Maaloev
Phone : +45 2527 7107, Fax : +45 4466 8856, Home +45 3874 5696
mail : peter.gram@xxxxxxxxxxxx - http://www.miracleas.dk
Upcoming events:
DBF2004 28-30 oktober 2004
Visit http://www.miracleas.dk fore news !
----------------------------------------------------------------
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:
