Re: How to apply archive logs on an cold backup

Frank:

I'm not heated or itchy. This isn't a matter of personalities. And it is disappointing that you attempt to make it so. If you must search for an adjective to apply to me personally, try 'passionate', because I am very passionate about not misleading newbies with dangerous and misleading advice constructed on nothing more than wild assumptions.

Bob posted simplistic and dangerous advice. He didn't like it when I called him on it. He can't justify his post on any technical grounds. I have talked about nothing but the technicalities of the matter since. Or at least tried to.

If you wish to get thoughtful and musing to the world in general, be my guest. But don't start speculating wildly about things that are not factual, in a post made physically in reply to one of mine, including an apparent accusation that I said things I never did, and expect me to be totally neutral about it, OK?

On that last specific issue, here is what you wrote:

Howard - why would this scenario never arise? > What scenario? Using backup controlfile? The scenario I tried to depict in the previous thread. > When did I ever say it never would arise?? Just reread your contributions to this thread

Here is what is in the fourth post (according to my newsreader) in this thread:

"The 'using backup controlfile' clause should only be specified if that's,
strangely enough, actually what you're using. Unless he restores his control files from backup, it would be lunacy to use the clause."

Here is my first post to Tim:
"ONLY if (1) you are in archivelog mode and (2) you have a backup
(temperature irrelevant) and (3) you lose EVERY copy of your control file and (4) you only have a binary backup of the control file

...only then do you issue the command 'recover database using backup controlfile'."

Here is another post I made in reply to Tim:
"Of course. You're wanting to perform recovery with a backup controlfile.
Strangely enough, that requires the use of the 'using backup controlfile' command."

Here is my penultimate reply to Tim:
"The correct recovery procedure for this sort of scenario is simply to
restore the control files and data files from your backup, and issue the command 'recover database until cancel using backup controlfile'. You could do an 'until time' if you know those sorts of detail, but usually it's an until cancel when lost redo logs are involved."

And I could go on.

I have repeatedly said, using backup controlfile is a valid command to use under the right circumstances.

So, I did what you asked, and re-read my contributions to this thread. Will you now therefore acknowledge that your hint that I said otherwise was inappropriate?

And as a specific explanation of what was wrong with your 'scenario', please see my response to Tim, because it is (as far as I can tell) exactly the same scenario he came up with.

"AFAICT, the OP wants to recover (PITR), but only happens
to have a cold backup, and all online, and off-line redo log files.
Now, you can restore your cold*) backup, and everything is honkey-dorey; however, you lost 2 days of data."

Now, your asterisk indicated that the OP would have restored
"Controlfile(s), online redolog files, datafiles, (S)PFILE(s),
directory structures, password file"

And if that's the case, all bets are off because he's just performed a wrong restoration. If he'd just restored the controlfile, the datafiles, the directory stuctures and so on, and NOT the online redo logs, then recovery is possible up to the point of the last log switch. No sweat, no worries, and no concerns whatsoever about the temperature of his backup.

And as I also said to Tim, could we please stop positing scenario after scenario, with rusty tapes and God knows what else? It rarely helps. The fundamentals never change, whatever the scenario. Don't backup your online logs; Restore what is broken; Construct a recovery command.

HJR Received on Fri Sep 03 2004 - 05:21:49 CDT