Comments on: Metalink Note on Datafile Recovery Will Corrupt Database

By: Alex Gorbachev

Alex Gorbachev — Thu, 04 Jun 2009 07:37:53 +0000

And if that doesn't work, Pythian can always help to recover data from a database that cannot be opened!

By: David Russell

David Russell — Mon, 19 May 2008 22:07:47 +0000

You’ve written a good article. I currently am working in a MS shop and only have two instances on Linux and do not plan on messing with them anytime soon. I still found this article interesting and the concepts should apply to other platforms, as well, even if not written up on Metalink.

What I would like to see, to complete the article, is for it not to be left in the state of “the critical step that’s missing is…” and see a definition of how to apply the critical step. From your article I cannot tell whether it is too late at this point to take those steps, or whether a totally different approach would be required because you “can’t get there from here”, or what?

By: Frits Hoogland

Frits Hoogland — Fri, 09 Nov 2007 15:35:31 +0000

Excellent follow up chris!

In the examples in my blog post (using the online redologfiles) I’ve already addressed the ‘open filedescriptors trap’ by shutting down the database during that example.

This just shows once again that careful planning of a recovery and knowing your stuff pays off, and that taking a shortcut could cost much more in the long run!

By: Christo Kutrovsky

Christo Kutrovsky — Fri, 09 Nov 2007 15:27:47 +0000

I’ve been thinking a bit more about this.

You would also need to do “begin backup” before the cat and “end backup” after the cat. Otherwise you may get fractured blocks.