Quantcast
Channel: SCN : All Content - SAP on SQL Server
Viewing all articles
Browse latest Browse all 837

SAP Migration - Import very slow - Source Oracle 11g - Target SQL Server 2005 S3

$
0
0

We are performing migration from oracle 11g to MSSQL server 2005 (Windows 2003/SAP 6.40).Our target system configuration is like below:

 

  • OS - Windows 2003 x64
  • DB - MSSQL server 2005 SP3 CU8
  • SAP – 6.40 Non Unicode
  • CPU - 8 processors (3.47ghz each)
  • RAM - 28 GB
  • Recovery Mode – Simple.
  • Temp DB 40gb DATA, 10gb Log (Since it is simple mode, it’s hardly getting use)
  • source system DB size - around 450 GB and package splitting was implemented.

 

While importing the export dump that came from source system, we are facing load performance issue's.

 

1) We have started 40 parallel processes and after import is started, CPU is getting 50% to 70% utilized.

  • Import Took around 6 days. It gets very slow on table S524 which is largest in source. After import Table S524 size in target was 600GB where in source it was only 100GB. Size of Rest of the table increase slightly, may be by 5%. Overall size of target DB is 950GB.

 

We have followed few SAP notes(1054852,1241751 e.tc..) and below are the settings for SQL server:

  • Minimum server memory - 13 GB
  • Maximum server memory - 13 GB
  • Index creation memory - 0
  • Minimum memory per query - 1024 kb
  • Maximum degree of parellelism - 0
  • Parallellism locks - 0
  • Cost threshold of parellism - 5
  • Enabled trace flags - 610,620,1117,1211,3917
  • Windows environment variable BCP_BATCH_SIZE = Not set
  • SQL log file size - 40 GB
  • Tempdb size - 50 GB

 

Are there any SQL server parameters need to be tuned to fix this import load time issue in both scenarios? Are there any  best practises for migrating to SQL server?

 

Export finished very quickly as expected in 10 hours.

 

Thanks..

Mamadi.


Viewing all articles
Browse latest Browse all 837

Trending Articles



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