Sapdata files are drastically increasing

[Former Member]

Hi friends,

Recently we have installed a new system that is ECC 6.0 on unix/oracle 10g, The SAPDATA files are drastically increasing they SAPDATA 1 and 2 are 100%. SAPDATA3 and 4 80%

Which are as below

99% /oracle/<SID>/sapdata2

78% /oracle/<SID>/sapdata3

80% /oracle/<SID>/sapdata4

100% /oracle/<SID>/sapdata1

Please help me how can i stop increasing this SAPDATA file increasing. I have scheduled the DB13 calender jobs.

But still the same.

Thanks

[Former Member]

Hello,

you should analyze which tablespace is using the datafiles that are growing:

select tablespace_name from dba_data_files

where file_name = '<path/datafile_name>';

And then you should also know which objects within that tablespace are using the most space, e.g:

set line 300

select owner, segment_name, segment_type,sum(bytes)/1024/1024 Bytes_MB

from dba_segments

where tablespace_name = '<TABLESPACE NAME>' and BYTES > (210241024*1024)

group by owner, segment_name, segment_type

See also these SAP Notes:

Note 16083 - Standard jobs, reorganization jobs

Note 974781 - Oracle database 10g: Internal maintenance jobs

Best regards

Baran

[Former Member]

After you installed the system, did you do a client copy to create your own client?

[sivaramakrishn]

One issue might be that SGEN is running for all components. However, this would be restricted to 1 or 2 filesystems only and not all.

The best way to find out is by using DB02 as others suggested. To help you better, we would need the tablespaces which grow, by how much they have grown and if possible the fastest growing tables and indexes.

Siva

[Former Member]

Hi,

What you report can only mean that the datafiles have AUTOEXTEND enabled and some activity in the system causes an extremely rapid growth of some of the files. Here's how I advise you to analyze this:

1) If you can afford to wait for a full day between two comparison points, then use DB02:

DB02 -> Space -> Segments -> Overview

Open tab "Top Growth"

This will show you the fastest growing segments (tables or indexes) along with their tablespace and growth rate

2) If the growth is critical and you cannot wait a whole day, then you can use SQLPLUS directly:

- log in as <sid>adm or ora<sid>

- commands:

sqlplus / as sysdba
spool /tmp/sizes1.lst
select segment_name,bytes/(1024*1024) "MBytes",tablespace_name
from dba_segments
where owner like 'SAP%'
order by 2 desc;
spool off
exit

- repeat the same command a few hours later but this time use 'spool /tmp/sizes2.lst'

- compare the largest segments between /tmp/sizes1.lst and /tmp/sizes2.lst

- once you know what the fastest growing segments are, you should check the activity in the SAP system (SM50 if currently running, SM37 for recent background jobs, ST03 or STAD for details on recent activity) to try and find out who is causing this rapid expansion.

Feel free to post the results to this forum, I'll do my best to help in the analysis.

Regards,

Mark

[Former Member]

This system is very newly built, not used by the users and not connected to the pinters nor there any spool requests. IS there any jdaily ob that i have to schedule in the SM37.

I have analysed DB02 but i don't find any relevant answer.

[Former Member]

What about the check with SQLPLUS? Something has to be the cause of this continuous growth.

Rgds,

Mark

[former_member204746]

you cannot really stop it from growing.

you need to analyze which datafiles are increasing and check if this is normal.

analyze with DB02.