|
Home > Archive > Oracle Server > July 2005 > queries on db corruption..
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
queries on db corruption..
|
|
|
| Hi ,
what options do we have when we find:
1)An object in sys or system schema is corrupt? -we are in noarchivelog
mode and take only logical backup.
as the object belons to sys/system , we can not drop and recreate
right? -so only way is to rebuild the entire database from the export
dump?
2) How do we proactively take steps to avoid corruption? -basically
what causes corruption or is known to cause it ( past experience etc) ?
Most thankful for help..
Nirav
| |
| sybrandb@yahoo.com 2005-07-25, 7:26 am |
| 1) only sys is not exported. And yes, your only option to rebuild the
entire database.
An excellent reason to switch to archivelog and implement proper
procedures
2) I didn't ran into corruption for the past 11 years, ie. once I left
the Vax-VMS world ;)
I'm not sure what kind of system you would need to experience
corruption.
But then *Rman* can detect corruption, and *Rman* should be used to
backup a database in *archivelog*. Those are the best proactive
measures you can take.
If you don't get any support for that from your management, it is
either time to say goodbye or wait until they blame you when disaster
strikes, and fire you.
Yours to choose!!
--
Sybrand Bakker
Senior Oracle DBA
| |
|
| 1) depends on what the object is, if its an index, Oracle MAY decide to
talk you through dropping it and recreating it using the data
dictionary creation scripts. If its a table, you dont have much option
other than to recreate the database from whatever backup you have.
2) You cant avoid corruption, you can only minimize the risk of loss of
data by having a proper DR plan.
However, regular checks for corruption can be done using the analyze
table command :-
ANALYZE TABLE/INDEX/CLUSTER ... VALIDATE STRUCTURE
Also dbverify on offline files which will scan all blocks in a datafile
and report on possible corruption :-
/home/oracle92$ dbv help=y
DBVERIFY: Release 9.2.0.4.0 - Production on Mon Jul 25 15:49:59 2005
Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.
Keyword Description (Default)
----------------------------------------------------
FILE File to Verify (NONE)
START Start Block (First Block of File)
END End Block (Last Block of File)
BLOCKSIZE Logical Block Size (2048)
LOGFILE Output Log (NONE)
FEEDBACK Display Progress (0)
PARFILE Parameter File (NONE)
USERID Username/Password (NONE)
SEGMENT_ID Segment ID (tsn.relfile.block) (NONE)
Phil.
| |
|
| Hi ,
Thank you experts..Unfortunately the clients insist on not changing
their ways so I don't have much option...I will try to be
"Disaster-Ready" and pray!
But here is one more question:
On Solaris 8 ,hosting oracle 9.2.0.6, can we run dbverfiy ONLINE?
Thanks again..
Ni...
| |
| sybrandb@yahoo.com 2005-07-26, 7:23 am |
| No. The option you have is to quit your current job.
Which I recommend, because sooner or later they will fire you.
Dbv can be run online.
But then again: corruption is very rare, even on Solaris.
You are fighting symptoms, and sooner or later disaster will strike,
following the 2 Laws of Murphy
The First Law of Murphy is: Everything that CAN go wrong, WILL go
wrong.
The Second Law of Murphy is: Murphy was an optimist.
--
Sybrand Bakker
Senior Oracle DBA
| |
|
|
|
| Thanks again gurus..Yes I do want to implement proper procedures...I
think while they don't agree for changing to archivelog mode, they will
agree to shut down the database once a week so that I can take cold
physical backup. Actually the clients drive the procedures so our
management really does not have much choice ...and clients are
technically very competent and they told me that server crash etc has
not happened in last 5/6 years so that is why they do not want to
change their ways..
So I will have daily full export backup as well as a cold backup once
in a week...some combination!!
The work environment is really smooth so I will continue here, with
prayers, cold & logical backups , and continued studies!!
N..
| |
| Joel Garry 2005-07-27, 1:24 pm |
|
nirav wrote:
> Thanks again gurus..Yes I do want to implement proper procedures...I
> think while they don't agree for changing to archivelog mode, they will
> agree to shut down the database once a week so that I can take cold
> physical backup. Actually the clients drive the procedures so our
> management really does not have much choice ...and clients are
> technically very competent and they told me that server crash etc has
>From what you have said, the clients are _not_ technically competent at
all. It is incumbent on your management to convince them to change
their ways while stroking their egos. With RMAN, you don't have to
shut down, and you don't have increased redo as with non-RMAN hot
backups.
> not happened in last 5/6 years so that is why they do not want to
> change their ways..
That is just stupid. Every so often we see posts like "we have raid,
we didn't have to worry about disks crashing. How can I recover my
data?" Proof that there are indeed stupid questions.
> So I will have daily full export backup as well as a cold backup once
> in a week...some combination!!
> The work environment is really smooth so I will continue here, with
> prayers, cold & logical backups , and continued studies!!
> N..
Well, if some data for the last couple of weeks is not too important to
lose, I guess that's ok then.
By the way, I once saw a place that diligently backed up their data
every day. They didn't have a good backup for at least 18 months,
though, since it was at least that long that it didn't fit on the tape,
which had errors to /dev/null.
jg
--
@home.com is bogus.
DJ: "[Over There] has a scary female private part."
SB: "I beg your pardon?"
DJ: "Rank! I meant rank!" - Steven Bochco interview heard on radio.
|
|
|
|
|