Home » RDBMS Server » Server Administration » LGWR
LGWR [message #141845] Wed, 12 October 2005 08:48 Go to next message
vens
Messages: 14
Registered: September 2005
Location: Russia
Junior Member

Hi,
The redo log file with "current" status is updated every 3 seconds as documentation tells us (as all controlfiles as well). In Unix we can easily see this with ls -l command. The timestamp is constantly being updated. But on Windows platform although lgwr is running OK the timestamp changes onle after commits, startups and so on - some "heavy" changes. Dir command displays so. Can you please explain it?
Re: LGWR [message #141853 is a reply to message #141845] Wed, 12 October 2005 09:07 Go to previous messageGo to next message
Mahesh Rajendran
Messages: 10708
Registered: March 2002
Location: oracleDocoVille
Senior Member
Account Moderator
>>The redo log file with "current" status is updated every 3 seconds as documentation tells
Could you please provide a link / URL to docset stating the above?
Re: LGWR [message #141857 is a reply to message #141845] Wed, 12 October 2005 09:21 Go to previous messageGo to next message
vens
Messages: 14
Registered: September 2005
Location: Russia
Junior Member

Sorry, i was n't exact -it is CKPT not LGWR but the question is still the same - on unix i see controlfiles and online redo getting new timestamp constantly, but not on windows

-----------------
http://download-west.oracle.com/docs/cd/B10501_01/server.920/a96524/c04space.htm#10136
--------------
"Control files also record information about checkpoints. Every three seconds, the checkpoint process (CKPT) records information in the control file about the checkpoint position in the online redo log. This information is used during database recovery to tell Oracle that all redo entries recorded before this point in the online redo log group are not necessary for database recovery; they were already written to the datafiles."
Re: LGWR [message #141861 is a reply to message #141857] Wed, 12 October 2005 09:44 Go to previous messageGo to next message
Mahesh Rajendran
Messages: 10708
Registered: March 2002
Location: oracleDocoVille
Senior Member
Account Moderator
did you check log_checkpoint_timeout?
Are log_checkpoint_interval / Log_Checkpoint_timeout differ in both the databases?
Re: LGWR [message #141863 is a reply to message #141845] Wed, 12 October 2005 09:48 Go to previous messageGo to next message
vens
Messages: 14
Registered: September 2005
Location: Russia
Junior Member

Absolutely the same (both instances)
Re: LGWR [message #141874 is a reply to message #141863] Wed, 12 October 2005 10:59 Go to previous messageGo to next message
Mahesh Rajendran
Messages: 10708
Registered: March 2002
Location: oracleDocoVille
Senior Member
Account Moderator
I dont have a windows instance around, that reproduces the case ( The one i have is heavily used).
I will try to simulate and keep posted.
and according to docset
http://strogoff.unex.es/oradoc/oracle10g/win.101/b10113/architec.htm
CKPT is optional and could crosschecked with this.
select * from v$bgprocess where paddr <> '00';

Thanks
Re: LGWR [message #141989 is a reply to message #141845] Thu, 13 October 2005 01:34 Go to previous messageGo to next message
vens
Messages: 14
Registered: September 2005
Location: Russia
Junior Member

Here it is (CKPT still present)
SQL> select name from v$bgprocess where paddr<>'00';

NAME

-----

PMON
DBW0
ARC0
ARC1
LGWR
CKPT
SMON
RECO

8 rows selected.
Re: LGWR [message #141999 is a reply to message #141989] Thu, 13 October 2005 02:16 Go to previous messageGo to next message
girish.rohini
Messages: 744
Registered: April 2005
Location: Delhi (India)
Senior Member
The case you are talking about is due to the difference in file handling in both unix & windows.
As for Oracle its a no issue.

--Girish
Re: LGWR [message #142000 is a reply to message #141845] Thu, 13 October 2005 02:18 Go to previous messageGo to next message
vens
Messages: 14
Registered: September 2005
Location: Russia
Junior Member

So it is quite normal for windows ?
Re: LGWR [message #142001 is a reply to message #142000] Thu, 13 October 2005 02:20 Go to previous messageGo to next message
Mahesh Rajendran
Messages: 10708
Registered: March 2002
Location: oracleDocoVille
Senior Member
Account Moderator
It supposed to be experienced in windows too.
Something is different in your case.
Re: LGWR [message #142008 is a reply to message #141845] Thu, 13 October 2005 02:31 Go to previous messageGo to next message
vens
Messages: 14
Registered: September 2005
Location: Russia
Junior Member

having only 2 windows based instances under my control, trying to find a reason for the described topic
Re: LGWR [message #142071 is a reply to message #141845] Thu, 13 October 2005 06:41 Go to previous message
vens
Messages: 14
Registered: September 2005
Location: Russia
Junior Member

I user File monitor utility from www.sysinternals.com - it monitors windows file activity. The write operations on controlfiles occure exactly once per 3 seconds. But operating system doesn't update timestamps however?
Previous Topic: ORA-12224: TNS:no listener
Next Topic: ora 9i passwword
Goto Forum:
  


Current Time: Sun Jan 26 11:37:53 CST 2025