Home Page for the TeradataForum
 

Archives of the TeradataForum

Message Posted: Mon, 15 Oct 2001 @ 14:02:34 GMT


     
  <Prev Next>   <<First <Prev Next> Last>>  


Subj:   Re: INSERT INTO big table
 
From:   Petr Horsky

Thanks a lot to all who responded.

* The two tables have the same PI, it is "almost" unique - in this case, would MULTISET also help?

* The AMP distribution of both tables seems ok, I am trying to check it in detail on hashrow level, but it takes time...

* The query runs under $H priority, no other queries run concurrently.

* We will work on COMPRESS. How could DEFAULT help in this case?

* We will add more HW soon (actually we will have to buy completely new box - we were told that it is not possible to expand our WM 4700 bought in 06/1999 as it is outdated - is this 99% scalability?)

Currently:

* HW: WM4700, two nodes, 24 AMPs, MP-RAS

* SW: V2R.03.00.03.42

* DBS Control Record - Performance Fields:
    1.  DictionaryCacheSize    = 128 (kilobytes)
    2.  DBSCacheCtrl           = TRUE
    3.  DBSCacheThr            = 10%
    4.  MaxParseTreeSegs       = 128
    5.  ReadAhead              = TRUE
    6.  StepsSegmentSize       = 1024 (kilobytes)
    7.  RedistBufSize          = 4 (kilobytes) (amp-level buffering only)
    8.  DisableSyncScan        = FALSE
    9.  SyncScanCacheThr       = 10%
   10.  HTMemAlloc             = 0%
   11.  SkewAllowance          = 75%

(Mr. Kohut, I hope this is what you meant by "internal performance options")

The average insert speed is 127 rows/node/sec or 10.6 rows/AMP/sec. OK or not?

Best regards

Petr Horsky



     
  <Prev Next>   <<First <Prev Next> Last>>  
 
 
 
 
 
 
 
 
  
  Top Home Privacy Feedback  
 
 
Copyright for the TeradataForum (TDATA-L), Manta BlueSky    
Copyright 2016 - All Rights Reserved    
Last Modified: 27 Dec 2016