Cluster table problem

[Former Member]

Dear experts,

I've a question regarding clust table.

When performing the step of "Create Cluster in Sender System", TDMS create table DMC_INDXCL in the tablespace with the biggest available freespace.

My question is:

1. Is it a temporary table? What time does this table been deleted? Is it been deleted in "Refresh Data Selection Cluster of Current Transfer"? Is there any other cluster table in sender system?

2. How much data will be filled? I mean if the productive database of size of 1TB, how much of the maximun data of this table?

3. I've found that, each package use the same name of DMC_INDXCL. Since the table will always been refreshed, if I'm running the different packages in parallel, is it possible that, the content of the previous package in cluster table will be refreshed in new package?

For example, I've got the PRD, QAS, DEV systems, sender system is PRD. The first package has setup the receive system as QAS, the next package set the receive system as DEV.

Shall I wait the until the first package to be finished?

Thanks, Jett

Edited by: Jett Jiang on Mar 18, 2009 6:01 AM

[Former Member · Accepted solution]

Hi Jett

Please find answers to your questions below -

>Is it a temporary table? What time does this table been deleted? Is it been deleted in "Refresh Data Selection Cluster of Current Transfer"? Is there any other cluster table in sender system?

The table is permanent but the data in the table is temporary. The other cluster table in the sender system is CNVMBTCLU.

>How much data will be filled? I mean if the productive database of size of 1TB, how much of the maximun data of this table?

Well - this depends . As storage of data in these clusters will alway lead to some compression - the assumtion is, that the total table size might be around to 10%-15% of used DB size for tables

>I've found that, each package use the same name of DMC_INDXCL. Since the table will always been refreshed, if I'm running the different packages in parallel, is it possible that, the content of the previous package in cluster table will be refreshed in new package?

For example, I've got the PRD, QAS, DEV systems, sender system is PRD. The first package has setup the receive system as QAS, the next package set the receive system as DEV.

Shall I wait the until the first package to be finished?

These tables have the intelligence build into them so as to avoid interference of data from different packages, so you do not need to wait for one package to finish to start the second package.

I hope these answers are of help.

Regards

Pankaj

[Former Member · Accepted solution]

Hi Pankaj,

Thank you so much for the reply! BTW, do you also have the experience in cramble on HR tables and customizing?

Thanks, Jett

[Former Member · Accepted solution]

Hi Jett

Please post your query regarding scrambling HR data in a new thread. We will try to answer the same

Regards

Pankaj.

[shouvik]

Hi Jett,

I know this is a bit late, but to simplify your analysis, what I do is I find out the space taken by the cluster table "DMC_INDXCL" before and after running Data Selection. To do that follow the following steps:

Go to transaction SE38 and execute the program "RSSPACECHECK"  or "RSTABLESIZE". Give "DMC_INDXCL" as the table name.

It will give you the total size of the cluster table.

Hope this helps.

Best Regards,

Shouvik