Hi Guys,

I am working on db2 luw version 8.1 on aix .


I have generated a reorgchk report for the tables in my database , For a few of them i could see some "*" for some of the tables .

To my understanding if for any formula for that table u see a "*" u can blindly do a reorg for it .

Am I right ? or is it that even if you find a "*" its not compulsory that reorg is required for that table , is there some additional criteria to check that .

Thanks
Sandeep

Hi,

http://public.dhe.ibm.com/software/d...ages_0508I.pdf
see section "Improving availability during table and index maintenance"

Cheers

I would suggest some reading about the command itself before acting on any message
try to understand what is going on
REORGCHK

Hi ,

Thanks for the link .

But it doesn't answer my question .

I still want to know if you see a "*" ,then should u chk any additional info or you can blindly do a reorg or not .

Right now my concern is not with the position of the "*" and which option to use with reorg (clean up etc ) .

If you want to, you can blindly reorg each table every day.

Or you can spend a few second and check if a reorg is necessary or not. That's why przytula_guy and i provided you the links.

you should check what the * stands for and analyze if the cause is important for you
in case it is for space reclamation - it might be you have space enough and this is a warning for you
if clustering is low - it might be important and I decide todo a reorg..
as indicated before : try to understand what is going on and how to act on this

Got it !!

Thanks
Sandeep

Depending on your table/index definitions, you may not be able to get rid of all '*'. For example, for the index cluster ratio. So always reorganizing a table if some '*' occurs somewhere is most likely the wrong approach.

This is my experience also (and other issues besides index cluster ratio). IMO the reorgchk program is a piece of junk and cannot be relied upon to automatically initiate reorgs. This is not a big concern if there is a defined maintenance window and one can do classic reorgs (offline), but is a problem for inplace (online) reorgs, especially since they run asynchronously, and one does not want to have more than few running at any given time (which is hard to control with asynchronous background processes).

The two main things you need to worry about for tables are clustering (if you have a clustering index and there have been a lot of inserts and deletes), and especially overflow pages. For indexes, badly fragmented indexes should be reorged if practical, although for an OLTP application you may not see a huge benefit.

am not sure this will help you
you can know the need of "reorg table" by checking reorg pending state of the table. if the value is greater than 3 execute reorg cmnd