Home » Infrastructure » Linux » I\O improvment
I\O improvment [message #424230] Thu, 01 October 2009 00:04 Go to next message
ursusca
Messages: 40
Registered: September 2009
Location: Toronto, ON
Member

Hi everybody,

I am interested in "noatime" mount parameter for ext2 and ext3 file systems.
Does anybody know if it would really improve I\O?

Thanks
Re: I\O improvment [message #424250 is a reply to message #424230] Thu, 01 October 2009 01:55 Go to previous messageGo to next message
tahpush
Messages: 961
Registered: August 2006
Location: Stockholm/Sweden
Senior Member

SD's, Journaling, and noatime/relatime
Re: I\O improvment [message #425967 is a reply to message #424230] Tue, 13 October 2009 07:46 Go to previous messageGo to next message
ursusca
Messages: 40
Registered: September 2009
Location: Toronto, ON
Member

I tried an experiment and I could not tell any significant difference between the performance with "noatime" mount option and without it.
Re: I\O improvment [message #425969 is a reply to message #425967] Tue, 13 October 2009 07:48 Go to previous messageGo to next message
Michel Cadot
Messages: 68624
Registered: March 2007
Location: Nanterre, France, http://...
Senior Member
Account Moderator
You should post this in Linux forum instead of an Oracle one.

Regards
Michel
Re: I\O improvment [message #425974 is a reply to message #425969] Tue, 13 October 2009 08:10 Go to previous messageGo to next message
ursusca
Messages: 40
Registered: September 2009
Location: Toronto, ON
Member

Michel Cadot wrote on Tue, 13 October 2009 16:48
You should post this in Linux forum instead of an Oracle one.

Regards
Michel


Hello,

I don't agree with you. Because operating system doesn't update the last access time for directories and files mounted under "noatime" mount point which translates into radically reduce total I/O. Linux offers this option to improve I/O for databases. That's why I tested it on RHEL5 with Oracle 10.2.0.4 but I didn't find any significant benefits of it.
Re: I\O improvment [message #425979 is a reply to message #425974] Tue, 13 October 2009 08:43 Go to previous messageGo to next message
Michel Cadot
Messages: 68624
Registered: March 2007
Location: Nanterre, France, http://...
Senior Member
Account Moderator
And in what this changes my answer?
I worked with Oracle since many years and I don't understand a word of what you say, this tends to prove this is not an Oracle question and you should better ask Linux expert than Oracle one.

Regards
Michel
Re: I\O improvment [message #426031 is a reply to message #425974] Tue, 13 October 2009 18:55 Go to previous messageGo to next message
ebrian
Messages: 2794
Registered: April 2006
Senior Member
Most all Linux tuning guides (and many Oracle tuning references) recommend setting this for applicable mount points and the link tahpush posted above does indicate one poster's performance gains, however YMMV based on your testing and disk setup.
Re: I\O improvment [message #426060 is a reply to message #426031] Wed, 14 October 2009 00:46 Go to previous message
ursusca
Messages: 40
Registered: September 2009
Location: Toronto, ON
Member

ebrian wrote on Wed, 14 October 2009 03:55
Most all Linux tuning guides (and many Oracle tuning references) recommend setting this for applicable mount points and the link tahpush posted above does indicate one poster's performance gains, however YMMV based on your testing and disk setup.


Hello ebrian

Thank you for attention to my question. I know that "noatime" mount option improves I/O by avoiding writes. But as I see people rarely care about the atime of an Oracle tablespace.
Previous Topic: can not start sqlplus from the shell
Next Topic: start Oracle forms from linux
Goto Forum:
  


Current Time: Thu Mar 28 06:50:42 CDT 2024