Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: W2000 connect / as sysdba problem

Re: W2000 connect / as sysdba problem

From: Howard J. Rogers <hjr_at_dizwell.com>
Date: Fri, 30 Apr 2004 21:52:24 +1000
Message-ID: <40923dff$0$442$afc38c87@news.optusnet.com.au>


Kenneth Koenraadt wrote:

> On Fri, 30 Apr 2004 18:03:11 +1000, "Howard J. Rogers"
> <hjr_at_dizwell.com> wrote:
> 
> 

>>Howard J. Rogers wrote:
>>
>>

[snip]

>>
>>Oh, by the way, just another test to make the point. Here's what my
>>server says:
>>
>>SQL> show parameter remote_login
>>
>>NAME TYPE VALUE
>>------------------------------------ ----------- ---------
>>remote_login_passwordfile string EXCLUSIVE
>>
>>Here's my sqlnet.ora on the *CLIENT* machine:
>>
>>SQLNET.AUTHENTICATION_SERVICES= (NTS)
>>
>>NAMES.DIRECTORY_PATH= (TNSNAMES)
>>
>>And here's the acid test:
>>
>>H:\>sqlplus "/@win92 as sysdba"
>>
>>SQL*Plus: Release 10.1.0.2.0 - Production on Fri Apr 30 18:00:36 2004
>>
>>Copyright (c) 1982, 2004, Oracle. All rights reserved.
>>
>>
>>Connected to:
>>Oracle9i Enterprise Edition Release 9.2.0.1.0 - Production
>>With the Partitioning, OLAP and Oracle Data Mining options
>>JServer Release 9.2.0.1.0 - Production
>>
>>So that's me finally managing to get a remote connection using O/S
>>authentication to a 9i database that's got R_L_P set to something other
>>than NONE.
>>
>>Still think the docs are correct?
>>
>>Regards
>>HJR

> 
> 
> 
> Your example does not prove anything.

Other than that your statement that "remote_login_passwordfile means "Ignore the password file, use O/S authentication"" is demonstrably false, you mean?

>  I have experienced myself  lots of time that I need to have R_L_L <>
> EXCLUSIVE" in order for "connect / as sysdba" to work, when I am
> logged on as a domain user being a member of ORA_DBA.
> 
> And  the Doc just confirms my observations, and I see no reason not to
> stick to it. 

The fact that your "observations" are demonstrably false doesn't have anything to do with it, I suppose.

> The fact that *you* might have been able to bypass it

"Bypass"??? By setting one init.ora parameter to the "wrong" value (according to you), having SQLNET.ORA set to a default value, and testing it????? You call that "bypassing" it??????

, possibly due to
> a W2K bug, is unusable to me. Sorry.

Oh come on. You said RLP had to be set to NONE for O/S authentication to work. I've just demonstrated that it doesn't. And it's not Win2K specific, because it does the same on Solaris (to name but one).

I don't need to "prove anything"... because the test results speak for themselves.

> BTW : How come that when you *think*

i.e, demonstrate.

>others are wrong, it's a "howler"

Because I've been aware of this as a documentation failing since version 8.0 was released. I apologise if you felt slighted by my calling it a howler. The bit about "utterly untrue", however, is demonstrably the case.

> and "utterly untrue". When YOU say something definetely incorrect (the > sqlnet.ora thing) it's : "Incidentally, I got it wrong".

Actually, it was "I was daft", but whatever: Check Google. I've described lots of my errors in terms akin to 'howlers'. But let's not get personal about it. You were wrong, and RLP does not need to be set to NONE to make remote O/S logins possible. Sorry, but I demonstrated it and you didn't. So the rest of the group can make up their own minds about who and what to believe.

> I suspect you get furious now, so I'll end the discussion here.
> 
> - Kenneth Koenraadt

Uh huh. When all else fails, and you are proved to be incorrect, take refuge in vague statements about your personal "experience" rather than demonstrated test cases; throw in some personal stuff to divert attention; and then cut-and-run and hope no-one will notice.

I thought better of you than that.

HJR Received on Fri Apr 30 2004 - 06:52:24 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US