Redo log switches in every 3 mins [message #567751] |
Fri, 05 October 2012 01:20 |
|
Dear All,
Today i noticed one problem with my database,my redologs switches in every 3mins,i also noticed there is no more transaction changes happening in database but still redo switches.
Fri Oct 05 06:10:05 2012
Thread 1 advanced to log sequence 79244
Current log# 2 seq# 79244 mem# 0: D:\ORADATA\ORACI\REDO02.LOG
Fri Oct 05 06:12:16 2012
Thread 1 advanced to log sequence 79245
Current log# 1 seq# 79245 mem# 0: D:\ORADATA\ORACI\REDO01.LOG
Fri Oct 05 06:14:28 2012
Thread 1 advanced to log sequence 79246
Current log# 3 seq# 79246 mem# 0: D:\ORADATA\ORACI\REDO03.LOG
Fri Oct 05 06:15:51 2012
Thread 1 advanced to log sequence 79247
Current log# 2 seq# 79247 mem# 0: D:\ORADATA\ORACI\REDO02.LOG
Fri Oct 05 06:16:59 2012
Thread 1 advanced to log sequence 79248
Current log# 1 seq# 79248 mem# 0: D:\ORADATA\ORACI\REDO01.LOG
Fri Oct 05 06:19:08 2012
Thread 1 advanced to log sequence 79249
Current log# 3 seq# 79249 mem# 0: D:\ORADATA\ORACI\REDO03.LOG
Fri Oct 05 06:21:22 2012
Can any one advice me why redo switch happening,any internal problem causes redo to switch please advice.
Regards,
Anil
|
|
|
|
Re: Redo log switches in every 3 mins [message #567767 is a reply to message #567752] |
Fri, 05 October 2012 03:26 |
|
Dear Michel,
"Quote"
i also noticed there is no more transaction changes happening in database
Kindly find what i did,
SQL> SELECT ADDR,XID AS "txn id", XIDUSN AS "undo seg", XIDSLOT AS "slot",
2 XIDSQN AS "seq", STATUS AS "txn status"
3 FROM V$TRANSACTION;
ADDR txn id undo seg slot seq txn status
-------- ---------------- ---------- ---------- ---------- ----------------
329DCBC8 02001700CB282400 2 23 2369739 ACTIVE
SQL> select sid,serial#,taddr,osuser from v$session where taddr='329DCBC8';
SID SERIAL# TADDR OSUSER
---------- ---------- -------- ------------------------------
380 8139 329DCBC8 atiacoh
SQL> select sql_text||' # '||hash_value
2 from v$sqltext
3 where (hash_value,address) in
4 (select sql_hash_value,sql_ADDRESS from
5 v$session where sid in (380))
6 order by piece;
no rows selected
there is only one transaction happening,i dont think this will cause my redo to full.
please find archive_lag_taget parameter.
SQL> show parameter archive_lag_target
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
archive_lag_target integer 0
SQL>
Regards,
Anil
|
|
|
Re: Redo log switches in every 3 mins [message #567768 is a reply to message #567767] |
Fri, 05 October 2012 03:34 |
|
Michel Cadot
Messages: 68718 Registered: March 2007 Location: Saint-Maur, France, https...
|
Senior Member Account Moderator |
|
|
Quote:there is only one transaction happening
Which does not mean it does not generate GB of redo/undo.
Check USED_UBLK of v$transaction.
In addition, you then see only the current on-going transactions and the history of them.
Regards
Michel
[Updated on: Fri, 05 October 2012 03:34] Report message to a moderator
|
|
|
Re: Redo log switches in every 3 mins [message #567797 is a reply to message #567768] |
Fri, 05 October 2012 07:30 |
|
Dear Michel,
Sorry for little delay in respose,i just monitored current trasaction can u please adivce below transaction really impact more redo switch based on USED_UBLK.
SQL> 1
1* select ADDR,STATUS,START_SCN,PRV_XID,recursive,USED_UBLK,XIDSQN from gv$transaction
SQL> /
ADDR STATUS START_SCN PRV_XID REC USED_UBLK XIDSQN
-------- ---------------- ---------- ---------------- --- ---------- ----------
329B6DD8 ACTIVE 5195055740 0000000000000000 NO 1 318666
329B7D7C ACTIVE 5195353662 0000000000000000 NO 1 284890
329DD104 ACTIVE 5193047179 0000000000000000 NO 9 1627239
329EE5F0 ACTIVE 5194799420 0000000000000000 NO 1 1456477
32A00A80 ACTIVE 5195368652 0000000000000000 NO 1 889994
32A24E74 ACTIVE 5195197115 0000000000000000 NO 2 2287717
32A26DAC ACTIVE 5192099121 0000000000000000 NO 4 1584660
32A6CBE4 ACTIVE 5195251441 0000000000000000 NO 1 54249
8 rows selected.
Also please note still redo switches in every 3 mins
Fri Oct 05 10:20:53 2012
Thread 1 advanced to log sequence 79357
Current log# 3 seq# 79357 mem# 0: D:\ORADATA\ORACI\REDO03.LOG
Fri Oct 05 10:23:14 2012
Thread 1 advanced to log sequence 79358
Current log# 2 seq# 79358 mem# 0: D:\ORADATA\ORACI\REDO02.LOG
Fri Oct 05 10:25:43 2012
Thread 1 advanced to log sequence 79359
Current log# 1 seq# 79359 mem# 0: D:\ORADATA\ORACI\REDO01.LOG
Fri Oct 05 10:28:10 2012
Thread 1 advanced to log sequence 79360
Current log# 3 seq# 79360 mem# 0: D:\ORADATA\ORACI\REDO03.LOG
Fri Oct 05 10:30:37 2012
Thread 1 advanced to log sequence 79361
Current log# 2 seq# 79361 mem# 0: D:\ORADATA\ORACI\REDO02.LOG
Fri Oct 05 10:33:08 2012
Thread 1 advanced to log sequence 79362
Current log# 1 seq# 79362 mem# 0: D:\ORADATA\ORACI\REDO01.LOG
Regards,
Anil
|
|
|
Re: Redo log switches in every 3 mins [message #567800 is a reply to message #567797] |
Fri, 05 October 2012 08:19 |
|
Michel Cadot
Messages: 68718 Registered: March 2007 Location: Saint-Maur, France, https...
|
Senior Member Account Moderator |
|
|
There you see you have many transactions and not just one.
I bet you have many and many small transactions.
You can use Log Miner and sum redo_length grouping by transaction.
Regards
Michel
|
|
|