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: fast_start_mttr_target value is ignored

Re: fast_start_mttr_target value is ignored

From: Howard J. Rogers <hjr_at_dizwell.com>
Date: Wed, 23 Jun 2004 06:32:01 +1000
Message-ID: <40d89724$0$25462$afc38c87@news.optusnet.com.au>

"Jack" <admin_at_stirlingonline.com> wrote in message news:7a80a637.0406221149.3df85be1_at_posting.google.com...
> Hi Howard;
>
> I didn't mean to offend you with any smart comments, but come on,
> sense of humor? We're not talking about saving lives here...
>
> I do appreciate very much the time you've taken to help me and I'll
> repeat that I'm certainly not interested in wasting your time or mine.
>
> As far as completely changing the mode of operation, I don't believe
> that the statement "create spfile from pfile" causes any of the
> parameter values to be changed, I could be wrong. But that seemed to
> me to be innocuous.

I'll take the trouble to reply, because it may stop someone else behaving so stupidly. I am not looking for this to become an on-going debate, however, since the facts speak for themselves.

This isn't about my sense of humour. It's not even about the fairly innocuous procedure of creating an spfile from a pfile. It's about you completely throwing a sequence of tests and actions that were designed, from the the time of your first post onwards, to find out what the problem was. Your first post made me think four things:

One: is the spfile being used instead of the init.ora? Two: is a value for one of the two parameters that mttr_target replaces causing the setting for mttr_target to be ignored? Three: are multiple init.ora's causing confusion? Four: Are there any other factors, such as licensing, bugs, known issues?

So I asked you about one, and we were able to tick that off. I asked you about two, and we were able to tick that off. We were about to tick off item three, in preparation for dealing with item four (which by the way we could have tested with the alter system command without an spfile being in place, because the parameter is dynamic and settable even with a boring old init.ora), but you decided to get "clever", and re-negotiate point one, *but without saying a word about it*. That *is* wasting my time. That is wasting anyone's time who has already mentally regarded the spfile issue as 'dealt with'.

It is impossible to help or support under those sort of circumstances, and that is not because I lack a sense of humour or for any other personality defect which you might want to start accusing me of suffering from. It is plain, common sense. You do not cause points which have already been established to be non-issues to become issues *again* by quietly deciding to do 'your own thing', and then not even mentioning it until asked. Yes, creating an spfile is innocuous... but *NOT* when the use or non-use of one has explicitly been asked about as part of the process of providing help.

Asking for *any* help or support here or indeed anywhere else actually involves an unwritten contract. The helpee promises not to withold information, to be accurate about the information s/he does supply, and to stop being 'clever', instead only doing what the helper tells him or her to do/type/perform. The helper likewise promises not to mislead, to use all best endeavours to achieve an outcome and so on. You broke your side of that bargain on two counts. One, see above. Two, see below.

> Looking in v$option I find:
>
> PARAMETER
> ----------------------------------------------------------------
> VALUE
> ----------------------------------------------------------------
>
> Fast-Start Fault Recovery
> FALSE
>
> Which I guess implies that I'm in fact not running Enterprise Edition.

You stated it as fact that you were. Had you not mentioned anything at all, or had you mentioned that you were not sure of the Enterprise Edition issue, step four of my plan would actually have become step one. On such minor details so much hangs.

> A select from v$version returns:
>
> SQL> select * from v$version;
>
> BANNER
> ----------------------------------------------------------------
> Oracle9i Release 9.2.0.1.0 - Production

Learn your banners, then:

SQL> select * from v$version;

BANNER



Oracle9i Enterprise Edition Release 9.2.0.1.0 - Production

If it doesn't say Enterprise Edition, don't assume it is. And certainly don't tell anyone it is when you don't know for certain one way or another.

A lot of people on this group have a killfile for dealing with problem posters. I don't have a killfile of quirky questioners, but if I did, you'd be added to it. Mistakes about the edition of Oracle you are using are perhaps understandable: not every DBA is involved in the purchasing and licensing of their particular Oracle installation, after all. But the spfile business is not. I hope you can perhaps see why. But if not, I really don't want to know about it.

HJR Received on Tue Jun 22 2004 - 15:32:01 CDT

Original text of this message

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