Dear all,

Here is something that happened to me this morning (11/30/04). I would
welcome comments from the database recovery experts on the list, and the
non-gurus like me might benefit in the future.

I had a database crash from a system failure. A power surge & subsequent
drop caused a disk controller to become flaky.

Upon replacement of the controller, the Oracle instance (9.2.0.1.0) would
mount, but not open. Open failed with an ORA-00600 failure. I looked up the
failure - after much fishing around for the oracle ORA-600 parameter analyzer
tool, and I also opened a TAR. The failure indicated a

This instance was running in NOARCHIVELOG mode, and recovering recent data
additions was not a priority. If I could bring it up at all, that would be
sufficient. I could easily determine what data had been lost.

The response from the TAR was speedy, but the first recommendation was to do
a backup (if I could), then recover from backup. While I was waiting for the
system running RMAN (which was also disabled by the power failure) to come
up, I did some checking around to see what I could determine about the state
of files in the db.

Two things I tried, here are the requests & complete responses:

#1. :SQL> Recover tablespace system;
:Media recovery complete.

A subsequent attempt to open failed.

#2. :SQL> recover database;
:Media recovery complete.

After this, an attempt to open was successful. Hooray!

Everything seems fine, but I'm wondering what I might have corrupted /
clobbered / destroyed. I'm also wondering why these two steps weren't in the
official "list of things to try" from Oracle's TAR reply.

TIA,
Tim





Oracle LazyDBA home page