|
Dear debian people,
I need to configure a application/database server to proof the potential of open source business intelligence. I want to use a ETL tool like the java based Talend or Pentaho Kettle or maybe the unix tools awk, sed, perl to transform 25 million records into 10 million records and loading them into a postgresql database server. Then I will use the java based Pentaho Report Designer to let users create reports or give the (power) users a query tool like pgadmin3 to access the database.
To design for maximum performance I've decided on the JFS or XFS filesystem, a raid 1 configuration where the OS en DATA will be on the same disks.
The hardware will be a workstation:
1 single-core 64 bit's CPU 2 SATA-II harddrives
Of course, the OS i've chosen will be Debian Lenny (stability, speed, apt-get, helpfull mailing list etc etc, the usual:-)
Questions 1) Can you show me the definitive debian lenny guide to create a software RAID-1 array from within the debian installer (I searched with google but wasn't that sucessfull and the manual is brief) The OS and data will be on the same disks as I don't have a spare OS disks only.
2) What OS tweaks do you advice, with my current setup plan (custom kernel, setting values, other filesystems like ext2)?
Thanks in advanced,
Dino |