Re: How can I get the BBED password?

  • From: "Slava Zayarny" <Slava.Zayarny@xxxxxxx>
  • To: <Oracle-L@xxxxxxxxxxxxx>
  • Date: Thu, 14 Apr 2005 09:55:45 -0400

Wolfgang,

I have recently been looking at what types of corruption RMAN notices
and how it reports them.  I used dd and BBED to experiment with =
intentional
block corruption in 9.2.0.6/HP-UX 11.11.

I have not had the time to conclude all the trials I want but BBED only =
marks=20
a block "FRACTURED" as seen in V$DATABASE_BLOCK_CORRUPTION
after RMAN's "BACKUP VALIDATE CHECK LOGICAL DATABASE" is run. =20

For this type of corruption RMAN does NOT fail with an error after =
backup=20
validation: it logs an entry to the alert log and inserts rows into=20
V$DATABASE_BLOCK_CORRUPTION and V$BACKUP_CORRUPTION.  Alert log would =
show
such intentionally fractured block in the alert log as:

        "Bad header found during backing up datafile"

without the usual "ORA-XXX".

RMAN behaved differently in a recent case where corrupted block was
reported in alert log as:

        "Bad header found during buffer read"

In that case RMAN failed with "ORA-19566: exceeded limit of 0 corrupt =
blocks for file".
It was also reported with "ORA-XXX" into the alert log.  Unfortunately, =
I don't have=20
details on how it was reported in V$DATABASE_BLOCK_CORRUPTION.

I think it's important to know what kind of corruption RMAN can and =
cannot
report without the need to add auto monitoring of the alert log beyond =
the standard
"ORA-XXX" strings.

What I want to figure out conclusively by simulation is how RMAN reports =
the=20
other four types of corruption (as seen in V$DATABASE_BLOCK_CORRUPTION):

ALL ZERO, CHECKSUM, CORRUPT, LOGICAL

I will share my findings when I'm done, but in the meantime, I would =
welcome
your comments if you have any.

Cheers,

Slava Zayarny
Oracle DBA

-----------------------------------------------

Date: Wed, 13 Apr 2005 12:43:24 -0600
From: Wolfgang Breitling <breitliw@xxxxxxxxxxxxx>
Subject: Re: How can I get the BBED password?

I have used it to deliberately corrupt a block to find out how RMAN=20
report block corruptions. Of course there are other ways to corrupt=20
blocks but it seemed like a good excuse to build and use bbed.

Guerra, Abraham J wrote:

> Hello Tanel,
>=20
> Have you used the utility?  If so, did you discover the password or =
=3D
> somebody else told it to you?
>=20
> Thanks.
>=20
> Abraham
>=20
--=20
Regards

Wolfgang Breitling
Centrex Consulting Corporation
www.centrexcc.com



--
//www.freelists.org/webpage/oracle-l

Other related posts: