Quantcast
Channel: SCN : Discussion List - SAP HANA Developer Center
Viewing all articles
Browse latest Browse all 6412

SQL Update, HANA performance

$
0
0

Hi,

 

As mentioned in this other discussion http://scn.sap.com/thread/3252533 I'm testing SAP HANA on AWS.

 

Currently I'm experiencing very slow performance on UPDATES, so slow that I'm quite sure i have something wrong but I can't figure what is going on.

 

I'm currently working on a COLUMN table like this, the primary key is (PROGR asc, JOB asc).

 

table.png

 

The tests where performed filling the table with 1M, 5M, and 100M records. Here's a snapshot of the content of the table:

 

tableCont.png

 

Through ODBC i connect a VB.NET application to my HANA developer instance on AWS.

 

The application spawns X threads, each thread creates his own connection/command/reader objects.

 

Each thread loops through the "JOBS" he's been assigned (eg, thread A works on records having JOB = 00000001 to JOB 00000050, thread B from 00000051 to 000000100 and so on. The number of "JOBS" The number of "JOBS" assigned to each thread is given by TOTAL_NUMBER_OF_RECORDS / NUMBER_OF_THREADS )

 

here's some pseudo code of the action deone by each thread.

 

 

FOR EACH JOB ID ASSIGNED

 

     FOR K=101 TO 200

 

          UPDATE THE RECORD WITH "JOB" = CURRENT_JOB_PROCESSED AND "PROGR" = K

          SET NUM1 = 1, NUM2 = 2, NUM3 = 3, NUM4 = 4, NUM5 = 5, NUM6 = 6, NUM7 = 7, NUM8 = 8, NUM9 = 9, NUM10 = 10

 

     NEXT

 

NEXT

 

The same thing is done on a MSSQL Server, which does the job faster.

 

I am sure that i have something not configured well because the gap beetween the times is huge.

 

I disabled savepoints and log on HANA just to take out disk issues on HANA's side, before running the update test i explicitly load the table in memory (just to be sure). I was reading some docs to understand the DELTA MERGING thing, if iI have to configure it somehow etc, i was thinking about ROW/TABLE locks and so on.

 

So, i have concurrent updates on the same table, each update works always on different records (the records to update are previously divided among the threads using the KEY of the table), the values updated are on colums that are not indexed nor in priamary key).

 

HANA is very fast on same testing but for SELECT queries, but on UPDATE something is going on.

 

Any ideas?

 

Thank you all,

david.

 

 

EDIT: the instances on AWS on which i'm using HANA and MSSQL are the same m2.4xlarge

 

Message was edited by: David Mandujano


Viewing all articles
Browse latest Browse all 6412

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>