PK and NOT NULL

Quote:

---

Originally posted by jmodic
I think it's time to raise a TAR.

---

Last thing i want to ckeck is REDO info when i issue

Create table ... Nologging, to see why the updated info NULL$ column of SYS.COL$ from 0 to 1 is not getting into Redo.

Not sure if i can get such info using Logminer or so.. Can we get such info Jurij?

Abhay.

Why would you want to go through all that hassle with Logminer? I'm sure you have some testing instance somewhere to play with. Set it to archlog mode, perform a backup and then create a table with nologging option. Then restore the backup, perform a media recovery and see what the outcome regarding that table is.

As I assume your database is 9i I would be curious to know why you haven't set your primary to FORCE LOGGING mode? How can you afford any nologging operation if you have stanby database configuration?

Quote:

---

Originally posted by jmodic
As I assume your database is 9i I would be curious to know why you haven't set your primary to FORCE LOGGING mode? How can you afford any nologging operation if you have stanby database configuration?

---

Yes DB is 9.x, but it was upgraded from 7.x to 8.x & now 9.2.0.4 ..

To tell you, we handle only DB performance/mantainence.. There is other team which handles Backup/Init Params.. Looks like they missed this option or may be unware..

Any ways I will suggest them to put Primary DB in Force Logging mode.

Thanks
Abhay.

Quote:

---

Originally posted by jmodic
Why would you want to go through all that hassle with Logminer? I'm sure you have some testing instance somewhere to play with. Set it to archlog mode, perform a backup and then create a table with nologging option. Then restore the backup, perform a media recovery and see what the outcome regarding that table is.

---

I checked on my local DB.. After performing media recovery for the SYSTEM datafile ( The one i restored from old backup ).. The "NOT NULL" on PK Cols were intact..

Now what, raise a TAR?

Abhay.

what is the out put of the following query shows?

select owner, segment_name, bytes/1024/1024 MBs from dba_segments
where segment_name in ('RLSCSHEV', 'PK_RLSCSHEV')
/

what has size to do with this, if after adding not null index_ffs is used :confused:

Quote:

---

Originally posted by pando
what has size to do with this, if after adding not null index_ffs is used :confused:

---

Probably pk index has grown larger than table segment.

Quote:

---

Originally posted by pando
if after adding not null index_ffs is used :confused:

---

yes, otimizer is using Index_FFS if added a explict NOT NULL to any one of the PK Col List.

Quote:

---

Originally posted by nagarjuna
Probably pk index has grown larger than table segment.

---

The issue is not really why Optimizer is not using index_ffs but why did not "NOT NULL" appear in our Stand By Server.

Abhay.