No more room in database .

Database
Fan

No more room in database .

HI All,

 

I am facing a no more room in the database error while taking backup. I have increased the perm space of the database. but this still throws the same error.

 

i tried changing the databases but still the issue is the same. as far as space is concerned we have ample space.

 

can anyone help me out here?

6 REPLIES
Junior Supporter

Re: No more room in database .

Are you doing online backups?


Abhishek Jadhav
Fan

Re: No more room in database .

No. normal table backups on the database. (offline)

Enthusiast

Re: No more room in database .

did you check the free space by AMP ? and the tables highly skewed

Enthusiast

Re: No more room in database .

I had same problem.

60gb free space

tale "a" uses 9gb

 

doing

insert into NewA select * from OldA

gives me the "no more room" space

Senior Apprentice

Re: No more room in database .

Hi,

Almost certainly your problem is due to skewing of data in this table and/or one of the others  in the database.

 

(Ignoring 'global space accounting' for now) With Teradata, the space limit is actually imposed 'per amp' although you specify it at the database level.

You create a database with perm=200GB on a 200 amp system. The real space limit is 1GB per AMP.

 

In your case if you have "60 gb free space" that may not be evenly distributed across all AMPs. Even if it is then the table that you're working on may not be evenly distributed.

 

Check the free space and the space used for your table 'per amp'.

 

Cheers,

Dave

 

Ward Analytics Ltd - information in motion
www: http://www.ward-analytics.com
Highlighted
Junior Contributor

Re: No more room in database .

First check skew at the database level using dbc.DiskSpaceV to see the actual free space by AMP.

Then check the skew of the target table.

If both source and target share the same PI you can simply query dbc.TableSizeV, otherwise do a "what if this is the new" PI query

SELECT MIN(c), AVG(c), MAX(c)
FROM
 (
   SELECT
      HASHAMP(HASHBUCKET(HASHROW(PI column(s) of target table))) AS vproc,
      COUNT(*) AS c
   FROM source_table
   GROUP BY 1
 ) AS dt