Hello , i run db2v9 express-c edition for linux 32 bits on a 4 xeon processors with 8Gb of RAM , my problem is on my enviroment the users produces huge delete from tables , i have indexes and a litle shell script which put database in quiesce mode , reorg tables , reorg indexes for tables and run runstats , but from what i seen in my db2diag.log when i reorg the table :
MESSAGE : ADM5541W Rebuilding index with IID "2" in object with ID "257" and
table space ID "2" on table "ZZZZZ .xxxx".

from this i understand that the index is droped and recreated ? in this case i assume that i don't need to reorg the indexes for the table ? or my assumption is wrong ?

my database is populated with data (about 15Gb ) the biggest table has on it 2 milions rows , in this case the reorg table will take about 3 hours .



Thank you

If you do an offline reorg of the table, the indexes will get reorged also.

3 hours is a very long time for 2 million rows. You should review your bufferpool sizes (if not using STMM) and perhaps specify a temp tablespace to do the reorg in (see reorg command in the doc).

i know that is a very long time , this was happening after i upgraded my memory from 4Gb to 8Gb (about 3 weeks ago) before that the reorg table last only 45 minutes , i have tried tempspace in reorg table command , no luck with that and i am using STMM

DB2 for Linux 32-bit can only use about 1.1 GB of memory per instance. I would set the instance memory maximum in the dbm config to 1 GB if you are using STMM, and make sure your bufferpools are no larger than 800 MB total for all active databases. I would recommend hardcoding the bufferpools instead of setting them to -2 (automatic).

Or you could upgrade to 64-bit, but then you only get 2GB of memory because it is DB2 Express-C.

i have downloaded a time ago the first version of db2 v9 express-c which has the limitation at 4GB of RAM with 2 processors and i have also the 64-bit version .... so i will try the 64-bit version