Prev: how to do a ldapsearch from a telnet session
Next: newbie question..droping BIN tables
From: zigzagdna on 23 Apr 2010 22:02
I am using Oracle 11.1.0.7.1 on HP UNIX 11.23. To my surprise, I find
that
a block has been corrupted. My export dump backups fail when reading
corrupt blocks.
I cannot repair from a backup, this system is not a production system,
no rman backups were taken.
I have read about DBMS_REAPIR and trying it. When I run
Check_object procedure:
dbms_repair.CHECK_OBJECT (
schema_name => 'TEST',
object_name => 'ASGNMT_RANGE',
partition_name => null,
object_type => dbms_repair.table_object,
repair_table_name => 'REPAIR_TEST',
flags => null,
, it does not populate anything in REPAIR_TEST table and errors with
following error:
declare
*
ERROR at line 1:
ORA-01578: ORACLE data block corrupted (file # 5, block # 258)
ORA-01110: data file 5: '/testprd/oradb2/oradata/xxxx/s_xxx_01_d1.dbf'
ORA-06512: at "SYS.DBMS_REPAIR", line 294
ORA-06512: at line 8

As a result when I run dbms_repair.fix_corrupt_blocks it does not fix
anything because there is nothing in REPAIR_TEST table.

From: vsevolod afanassiev on 25 Apr 2010 05:40
Never used DBMS_REPAIR so can't help you with it. However it may be
more important to determine what caused this issue and how to prevent
it from happening again, especially in Prod environment. Did you get
any ORA-00600, ORA-07445 error? Was there are hardware issue? I/O
errors? Check UNIX log. What happens if you try to run dbv utility for
this file? Or copy it to /dev/null using 'cp' command?
From: John Hurley on 25 Apr 2010 06:06
On Apr 23, 10:02 pm, zigzagdna <zigzag...(a)yahoo.com> wrote:

snip

> I am using Oracle 11.1.0.7.1 on HP UNIX 11.23. To my surprise, I find
> that
> a block has been corrupted. My export dump backups fail when reading
> corrupt blocks.
> I cannot repair from a backup, this system is not a production system,
> no rman backups were taken.

Can you drop the tablespace and recreate it?

Can you run rman backup to disk and at least get rman to tell you how
many bad blocks you have?

From: zigzagdna on 25 Apr 2010 11:55
On Apr 25, 5:40 am, vsevolod afanassiev
<vsevolod.afanass...(a)gmail.com> wrote:
> Never used DBMS_REPAIR so can't help you with it. However it may be
> more important to determine what caused this issue and how to prevent
> it from happening again, especially in Prod environment. Did you get
> any ORA-00600, ORA-07445 error? Was there are hardware issue? I/O
> errors? Check UNIX log. What happens if you try to run dbv utility for
> this file? Or copy it to /dev/null using 'cp' command?

No clue what cused it, may be Oracle 11g is not stable Never happend
to me in previous version databases.
From: zigzagdna on 25 Apr 2010 11:57
On Apr 25, 6:06 am, John Hurley <hurleyjo...(a)yahoo.com> wrote:
> On Apr 23, 10:02 pm, zigzagdna <zigzag...(a)yahoo.com> wrote:
>
> snip
>
> > I am using Oracle 11.1.0.7.1 on HP UNIX 11.23. To my surprise, I find
> > that
> > a block has been corrupted. My export dump backups fail when reading
> > corrupt blocks.
> > I cannot repair from a backup, this system is not a production system,
> > no rman backups were taken.
>
> Can you drop the tablespace and recreate it?
>
> Can you run rman backup to disk and at least get rman to tell you how
> many bad blocks you have?

Only one block is corrupted. Oracle 11g provdies various health
monitors and only one block is shown as corrupted. My export dump
fails for one table. One can reterive dthat table from a prevous
export backup, but that will require some loss in data. I am trying
to figure what is the value of dbms_repair.
 |  Next  |  Last
Pages: 1 2
Prev: how to do a ldapsearch from a telnet session
Next: newbie question..droping BIN tables