Segment size

Hi i had a table with 45 lakhs of records. i had a trigger for updation. i had written an update query which updates 25 lakhs records. but i terminated the session while update is going on. i deleted what ever rows got updated in the trigger table. but now when i check in dba_segments my trigger table is showing aroung 2 GB size even though it has only 60,000 records.
what may be the problem? should i use alter tablespace <tablespace_name> coalecse; command?
plz guide me
thanks in advance.

The easiest way to clear up deleted block is using "MOVE" clause.
Syntax: ALTER TABLE tab_name MOVE options
But keep in mind that you need free spaces where the table moves.
For instance:
SQL> select bytes from dba_segments where segment_name = 'REDO_TEST';
BYTES
1048576
SQL> delete from redo_test;
SQL> commit;
SQL> select bytes from dba_segments where segment_name = 'REDO_TEST';
BYTES
1048576
SQL> alter table redo_test move;
SQL> select bytes from dba_segments where segment_name = 'REDO_TEST';
BYTES
65536

Similar Messages

Java NIO - TCP segment size abnormally low

Hi !
After noticing a weird behaviour on our Linux production server for code that works perfectly on my Windows dev box, I used tcpdump to sniff the packets that are actually sent to our clients.
The code I use to write the data is as simple as :
// using NIO - buffer is at most 135 bytes long
channel.write(buffer);
if (buffer.hasRemaining()) {
// this never happens
}When the buffer is 135 bytes long, this systematically results in two TCP segments being sent : one containing 127 bytes of data, the other one containing 8 bytes of data.
Our client is an embedded system which is poorly implemented and handles TCP packets as whole application messages, which means that the remaining 8 bytes end up being ignored and the operation fails.
I googled it a bit, but couldn't find any info about the possible culprit (buffer sizes and default max tcp segment sizes are of course way larger that 127 bytes !)
Any ideas ?
Thanks !

NB the fragmentation could also be happening in any intermediate router.
All I can suggest is that you set the TCP send buffer size to some multiple of the desired segment size, or maybe just set it very large, like 64k-1, so that you can reduce its effect on segmentation.

"Limit Capture/Export File Segment Size To" doesn't work

I put the limit to 4000 MB before i startet capturing HD-video, but it didn't work. Several of my files are bigger than 4 GB. This is a problem since I use an online back up service that has 5 GB as the maximum file limit. Any suggestion to fix this problem is highly appreciated.

I believe, although I am not 100% sure, that the "Limit Capture/Export File Segment Size To" does not apply to Log and Transfer, only Log and Capture.
Since Log and Capture works with tape sources, when the limit is hit, the tape can be stopped, pre-rolled and started again to create another clip
In the case of Log and Transfer, it is ingesting and transcoding existing files; the clip lengths (and therefore size) are already determined by the source clip.
If you are working with very lengthy clips, you may want to set in and out points to divide the clips into smaller segments.
MtD

Segment size problem

I use apache HttpClient upload xml file to a server, and it is https. It works for some files, but for some of them i can not even get response, and use snoop to check networking, i found out, for the bad cases, there is one frame shows "unreassembled packet" and the checksum is INCORRECT, the size of the frame is 1514, the total length in Internet Protocol is 1500, but in the beginning of connection it shows MSS=1460,
so who can explain the realtionship between these numbers, and where i can configure or control the segment size when uploading, or it is impossible.

yes, 1514 is right, and the checksum incorrect is beacuse i am using snoop right on my server, the checksum has not been calculated by NIC, it is always 0, now i noticed the problem happens when the segment reachs 1460, and then, this packet is "unreassambled". So, i am wondering where is the problem, SSL, or system configure, or somthing else. By the way, my server is solaris 10 , and the peer, i am not soure, IBM, but i think they may use proxy server.

Segment Size when Packaging

For a while now, when packaging material for our intranet I
use a custom setting that works best with our data transfer rates.
I am getting ready to put some courseware on our outside server for
other employees to access when they are offsite. My question is
should I use 56 kbs Modem or DSL/Cable Modem? If I use the higher
setting, how would it impact 56 kbs users? If I use lower settings,
would 56 kbs users notice any difference? Would cable modem users
see slower performance when packaged at lower settings?
Anyone have some experience in testing these settings?
Thanks!
Steve

> Anyone have some experience in testing these settings?
I always go for the larger segment size, for a number of
reasons, including:
- The browser has to make fewer calls for segments - each
request has a
certain overhead; latency waiting for the request to be
answered etc. Take
the extreme and try 1k segments and see how bad it can be.
- Larger segments may cause what seem like longer delays on a
single
request, but they download more code, so there are fewer
requests. The user
perception can be that the course is more usable.
- If you have any large graphics or other internal content,
they won't be
split up anyway, so many segments won't ever be so small as
16k.
Steve
EuroTAAC eLearning 2007
http://www.eurotaac.com
Adobe Community Expert: Authorware, Flash Mobile and Devices
My blog -
http://stevehoward.blogspot.com/
Authorware tips -
http://www.tomorrows-key.com

Average segment size of iDoc?

We want to integrate 2 systems by using iDoc across WAN. So far we know the iDoc volumes and average iDoc segments, for example 6000 SO iDoc/hour and 250 segments of each iDoc. My understanding is that we need to know the average iDoc segment size to performance network sizing.
I know the size varies from case to case. But is there a BEST PRACTICE value we can use?
TIA
James

Hi Jinquin,
>>I happen to find something interesting at http://msg-ibexi.wikispaces.com/message/view/home/10780864, "Maximun size of IDOC segment is 1000 bytes", is it true?
Yes
SDATA contains segment data and it is 1000 character long check here
http://help.sap.com/saphelp_47x200/helpdata/en/1a/0e3842539911d1898b0000e8322d00/content.htm
Also see the last statement in the link
http://help.sap.com/saphelp_nw04/Helpdata/EN/78/21740f51ce11d189570000e829fbbd/frameset.htm
Regards
suraj

Why Argument$ segment size is increased

Hello everyone,
I have imported dump file having size 800mb and after completing importing given file my database size became more than 10 gb and in that arguement$ segment occupied more than 5 gb.
so, I would like to know
1.why it worked in this way ..
2. As this is system tablespace's segment so ultimately System tablespace increased .. hence is there any way to reduce system TS.
below query result for ur reference..
Thanks in adavance
select owner,segment_name,segment_type ,bytes/(1024*1024) size_m from dba_segments where tablespace_name = 'SYSTEM' and bytes/(1024*1024) > 2 order by size_m desc;
OWNER SEGMENT_NAME SEGMENT_TYPE SIZE_M
SYS ARGUMENT$ TABLE 5431
SYS I_ARGUMENT1 INDEX 4366
SYS I_ARGUMENT2 INDEX 2453

Reducing system tablespace this will help
Shrink/Reduce System Tablespace!
System Tablespace have grown too large!
http://asktom.oracle.com/pls/asktom/f?p=100:11:0::::P11_QUESTION_ID:153612348067

DB Size is reduced (LOB segment size) after Migration from Win To Linux

Hi Friends,
We have migrated Oracle 11.2.0.1 Database (64 bit) from Windows (64 Bit) To Linux (64 Bit) using RMAN convert database at target.
After Migration i could see the size of LOB Segment is very less in targe so as the DB Size.
Is it baecuas the conversion extracts only the used segments from the source datafilles (or) am i losing some data?
SQL> DECLARE
2    db_ready BOOLEAN;
3 BEGIN
4    db_ready :=
5         DBMS_TDB.CHECK_DB('Linux IA (64-bit)',DBMS_TDB.SKIP_READONLY);
6 END;
7 /
PL/SQL procedure successfully completed.
SQL> DECLARE
2       external BOOLEAN;
3 BEGIN
4      /* value of external is ignored, but with SERVEROUTPUT set to ON
5       * dbms_tdb.check_external displays report of external objects
6       * on console */
7      external := DBMS_TDB.CHECK_EXTERNAL;
8 END;
9 /
PL/SQL procedure successfully completed.
Regards,
DB

>Is it baecuas the conversion extracts only the used segments from the source datafilles (or) am i losing some data?
I suspect that the source DB has many LOB rows deleted.

Raw format segment size

Hi! In raw format stream data are divided into many files named "segments". How can I configure the size of this files? I want to make them twice smaller.

You can use MaxFlushSize(Server.xml - Root/Server/ResourceLimits/RecBuffer/MaxFlushSize) configuration in Server.xml to tweak the size of segments which gets created. Minimum size is 32KB.

Maximum data segment size for a process

Hello
we have an issue in our proxy server with the size of a process.
the process grow in size and when it reaches 4Gbytes, the process stop with an error that it cannot allocate memory, we check and there's plently of swap left,. all the ulimit settings in solaris 10 are set to 'unlimited'
is there a way i can see or determine, bypass configure what is the maximum process size. on a 64bit solaris.
thank you in advance
Mario G.

It sounds like you're running a 32-bit app. You need a 64-bit application for it to be able to utilize more than 4 GB of RAM.

How to estimate a rollback segment size for a big transaction?

Oracle 8.1.5. Suppose I have the following statement:
update table_a set col_a = rownum where col_b = 'codeA'
The statement affect 900,000 rows. Type of col_a - float(126).

you can have RBS size same as of your data tablespace
(eg) you have 10 GB tablespace allocated for data
then same 10GB more less should be comfortable

LOB segment size is 2 times bigger than the real data

here's an interesting test:
1. I created a tablespace called "smallblock" with 2K blocksize
2. I created a table with a CLOB type field and specified the smallblock tablespace as a storage for the LOB segment:
SCOTT@andrkydb> create table t1 (i int, b clob) lob (b) store as
t1_lob (chunk 2K disable storage in row tablespace smallblock);
3. I insert data into the table, using a bit less than 2K of data for the clob type column:
SCOTT@andrkydb> begin
2 for i in 1..1000 loop
3 insert into t1 values (mod(i,5), rpad('*',2000,'*'));
4 end loop;
5 end;
6 /
4. Now I can see that I have an average of 2000 bytes for each lob item:
SCOTT@andrkydb> select avg(dbms_lob.getlength(b)) from t1;
AVG(DBMS_LOB.GETLENGTH(B))
2000
and that all together they take up:
SCOTT@andrkydb> select sum(dbms_lob.getlength(b)) from t1;
SUM(DBMS_LOB.GETLENGTH(B))
2000000
But when I take a look at how much is the LOB segment actually taking, I get a result, which is being a total mystery to me:
SCOTT@andrkydb> select bytes from dba_segments where segment_name = 'T1_LOB';
BYTES
5242880
What am I missing? Why is LOB segment is being ~2 times bigger than it is required by the data?
I am on 10.2.0.3 EE, Solaris 5.10 sparc 64bit.
Message was edited by:
Andrei Kübar

thanks for the link, it is good to know such thing is possible. Although I don't really see how can it help me..
But you know, you were right regarding the smaller data amounts. I have tested with 1800 bytes of data and in this case it does fit just right.
But this means that there is 248 bytes wasted (from my, as developer, point of view) per block! But if there is such an overhead, then I must be able to estimate it when designing the data structures. And I don't see anywhere in the docs a single word about such thing.
Moreover, if you use NCLOB type, then only 990 bytes fits into a single 2K chunk. So the overhead might become really huge when you go over to gigabyte amounts...
I have a LOB segment for a nclob type field in a production database, which is 5GB large and it contains only 2,2GB of real data. There is no "deleted" rows in it, I know because I have rebuilt it. So this looks like a total waste of disk space... I must say, I'm quite disappointed with this.
- Andrei

Avg_row_len Vs. Segment size

Hi
In our Production DB (size 650 GB), I am closely monitoring the
space growth pattern for the past one year.
Last week I observed that in some of the tables
avg_row_len*num_rows = 50 MB
and sum(bytes)/1024/1024 from user_segment where
segment_name ='<table_name>'
This gives app 600 MB.
Now my questions are :
1) What does this mean ?
2) Is there lot of free space available ?
3) Will this affect when going for full table scan ?
4) How to make it smaller considering the
actual size of the table ?
Please give me the solid solution which can be
practically implemented in production DB (No
academic, theoritical answers please)
Regards
S.D. RAVICHANDRAN
Verizon.

Hi Guys
I got answer after raising SR.
Oracle advised to use move command to optimize the space.
The bottomline is we should have enuf space. Say size of the
table+10% is sufficient.
Hope this will help other DBAs.
S.D. RAVICHANDRAN
Verizon.

Truncate log and device/segment size

If I have a ASE database, it's log is never be truncated, the log become very large. I do have backup every day, but backup did not reduce the size of log.
If I turncate the log today. the space on device for log will be released. Is it possible to reduce the size of the device for log on live production?
Also if log never be truncated, does it impact the ASE performance?

Hi Kent,
Refer to the discussion in thread http://scn.sap.com/thread/3467480
Hope this helps.
Regards,
Deepak Kori

Collect segment size history in 10g

Hi,
In order to know how much each segments and tablespace growth each month ,
i have a job that queering dba_segments and dba_data_files and save the results in a database tables.
Does 10g collect this information automatically , or should i continue collect this kind on statistics by my self ?
Thanks.

You can try to use the DBMS_SPACE package. For example:
SQL> column timepoint format a30
SQL> select * from
2 table(dbms_space.OBJECT_GROWTH_TREND(
3          'SYS', 'OBJ$', 'TABLE', null, to_timestamp('14-JUL-2008','DD-MON-YYYY'), null, numtodsinterval(30, 'DAY') ));
TIMEPOINT                      SPACE_USAGE SPACE_ALLOC QUALITY
14-JUL-08 12.00.00.000000 AM       1318466     2097152 INTERPOLATED
13-AUG-08 12.00.00.000000 AM       1318466     2097152 GOOD

Segment size

Similar Messages

Maybe you are looking for