-
Hi DBA's
I have a database which is not set to archive log. Every one hour it is generating a redolog file approximately. I checked my init.ora parameters and they are set as:: log_checkpoint_interval = 999999999
log_checkpoint_timeout = 0
my svrmgr session is:
SVRMGR> connect internal
Connected.
SVRMGR> archive log list
Database log mode No Archive Mode
Automatic archival Disabled
Archive destination /u01/app/oracle/product/8.1.6/dbs/arch
Oldest online log sequence 17348
Current log sequence 17350
my bdump shows:
$ pwd
/u01/app/oracle/admin/nbs/bdump
$ tail -15 alert_nbs.log
Wed Nov 29 11:17:54 2000
Thread 1 advanced to log sequence 17346
Current log# 3 seq# 17346 mem# 0: /u02/oradata/nbs/redo03.log
Wed Nov 29 12:14:00 2000
Thread 1 advanced to log sequence 17347
Current log# 1 seq# 17347 mem# 0: /u02/oradata/nbs/redo01.log
Wed Nov 29 13:10:06 2000
Thread 1 advanced to log sequence 17348
Current log# 2 seq# 17348 mem# 0: /u02/oradata/nbs/redo02.log
Wed Nov 29 14:06:10 2000
Thread 1 advanced to log sequence 17349
Current log# 3 seq# 17349 mem# 0: /u02/oradata/nbs/redo03.log
Wed Nov 29 15:02:18 2000
Thread 1 advanced to log sequence 17350
Current log# 1 seq# 17350 mem# 0: /u02/oradata/nbs/redo01.log
Can you please justify:::
Thanks
-
-
Oh.. My question is that, why is it creating these files at every one hour?? Please refer to my parameters of init.ora.
Thanks..
-
It is not generating archived log file.
When ever log switch occurs, the information is recorded in the alert.log fle.
Do not worry.
-
Sorry TamilS::
I mistyped my mind.. I am sure they are not creating any files. It was a dumb question.. My concern is why is it switching logs that frequently when the 2 parameters are not indicating the switch.
log_checkpoint_interval = 999999999
log_checkpoint_timeout = 0
Thx..
-
Those parameters control the FREQUENCY of the CHECKPOINT process{CKPT} A checkpoint will occur at redo- log switch... these parameters are important if you redo-logs are large.... more frequent checkpointing reduces recovery time....
-
In your case the log switch occurs every one hour. That means either your redo log file is big or less number of transaction takes place in an hour.
Oracle writes transaction info sequentially in redo log file. When the crrent one fills, it starts writing into next one.
-
Thats true..
I guess I have to be more clear.. Let me try again..
My redolog files are 5 mb each and there is no activity in the databse since morning. But everyhour a log switch is happening.
log_checkpoint_interval = 999999999
log_checkpoint_timeout = 0
With the above 2 params, how can I justify a log switch every one hour when no activity is going on in the DB..
And Tamil, if you get a chance can you also look closer into another question I posted about error-0131 Insufficient Privileges when trying to access db thru connect internal.
Thx
-
I think you got the concept confused, checkpoint doesnt force a log switch but I think you've got that concept the other way round so those 2 parameters you are mentioning will not force any log switch even they are set very low.
Now you ask why there are logs even there arent any activity, I dont know why this happens but the database I have administer so far I havent faced any kind of these problems although these problem some people have mentioned already in this forum and the suggestions were either this is kind of bug or there were really things going.
Anyway those 2 parameters doesnt force log switch!
-
There is either some activity in your database or somebody is manually switching the logfile by typing "alter system switch logfile"
