Re: 2 GB myth

Igor Laletin wrote:
> "Howard J. Rogers" <hjr_at_dizwell.com> wrote in message news:<41aad392$0$9113$afc38c87_at_news.optusnet.com.au>...
>
>>>You started with agruing with something I didn't say.
>>
>>Incorrect. I pointed out that saying 'recovery is not affected by the
>>number of disk files, especially with today's hardware' misses the point
>>entirely.
>
>
> Sometimes you should actually read before answering.

All times, you should try and learn English before posting in it.

>It was "bigger
> datafiles don't make recovery much slower. Not with the current
> hardware anyway". It's different from your "is not affected".

And the difference in meaning, outcome and implication of your phraseology versus mine is...???

Twist away, Igor. You dug yourself in, I have no objection to you attempting to dig yourself out.

The rest of the group, if they're even remotely interested, can see precisely what you're doing.

>>And is in any case not true when >1 file needs restoring.
>
>
> I won't even point it was _datafile_ recovery.

Almost every form of recovery is a data file recovery, Igor.

And, for the record, there are complete and incomplete recoveries. There's no main heading in the documentation for a "data file recovery".

But if what you are feebly attempting to say is, "I assumed only one file was involved in the recovery", then you assumed wrong, which is not a particular surprise in your case.

You might just cast your mind back to the original poster (remember him?) who wasn't actually asking about recovery at all. So on what basis   you can state "it was" anything at all is utterly mystifying. But then, that's no particular surprise in your case, either.

> As for many datafiles,
> there is some difference for small number of datafiles (especially
> from different tablespaces). The more datafiles, the less difference
> obviously.

Only obviously in your head. In the real world, we find out what the specific situation is for individual setups before we make sweeping statements like that.

Oh... but I forgot... sweeping generalisations are your forte.

> Now add time to place tapes, rewind to the right spot to
> start streaming, going through redo logs and you'll probably find
> larger file doesn't cost you that much.

"Probably". "Obviously". "Much". Weasel words, Igor. Weasel words.

Why don't you just admit it: what you originally wrote was a silly generalisation?

Otherwise, I have nothing further to add. Received on Tue Nov 30 2004 - 01:59:04 CST