IDS Forum

Re: onstat -g iof

Posted By: John Miller iii
Date: Saturday, 29 May 2010, at 5:47 a.m.

In Response To: Re: onstat -g iof (Fernando Nunes)

In short the server times every single I/O operation performed along with
how many I/O operations
have occurred by device. The reports average time per device is just the
time for all I/O operations
divide by the number of operations. While I understand that this can be
influenced by OS caches and
other OS and hardware performance aids. This is the true response time
that Informix is seeing.

This is formation is calculated since the server booted or an onstat -z
whichever is more recent.

John F. Miller III
STSM, Support Architect
miller3@us.ibm.com
503-578-5645
IBM Informix Dynamic Server (IDS)

ids-bounces@iiug.org wrote on 05/27/2010 04:20:58 PM:

> [image removed]
>
> Re: onstat -g iof [20250]
>
> Fernando Nunes
>
> to:
>
> ids
>
> 05/27/2010 04:21 PM
>
> Sent by:
>
> ids-bounces@iiug.org
>
> Please respond to ids
>
> Which is what I understand from the manual, although it's not very
explicit.
> Again, this is I/O operations, and it can be greatly influenced by
hardware
> caches.
>
> Regards.
>
> On Thu, May 27, 2010 at 11:49 PM, Art Kagel <art.kagel@gmail.com> wrote:
>
> > Someone from the Labs (maybe Scott Lashley, maybe John Miller III) told
me
> > once that the io/s is calculated as a rolling value that's only updated

> > when
> > there is actual IO being performed. The exact mechanics of how the
engine
> > implements that calculation I don't know, but it would be something
like
> > storing the total time spent performing IO divided in to the number of
IOs
> > not elapsed clock time since startup or since stats were zero'd.
> >
> > Art
> >
> > Art S. Kagel
> > Advanced DataTools (www.advancedatatools.com)
> > IIUG Board of Directors (art@iiug.org)
> >
> > Disclaimer: Please keep in mind that my own opinions are my own
opinions
> > and
> > do not reflect on my employer, Advanced DataTools, the IIUG, nor any
other
> > organization with which I am associated either explicitly, implicitly,
or
> > by
> > inference. Neither do those opinions reflect those of other individuals

> > affiliated with any entity with which I am affiliated nor those of the
> > entities themselves.
> >
> > On Thu, May 27, 2010 at 5:26 PM, ROBERT CLOW <robert@clow.biz> wrote:
> >
> > > Thanks again but I do not believe that the manual is correct
> > >
> > > onstat taken at 04:00
> > >
> > > 32 data_part_02_02 3899850752 952112 52940800 12925 660.4
> > >
> > > op type count avg. time
> > >
> > > seeks 0 N/A
> > >
> > > reads 0 N/A
> > >
> > > writes 0 N/A
> > >
> > > kaio_reads 259790 0.0013
> > >
> > > kaio_writes 921 0.0605
> > >
> > > And at 08:00
> > >
> > > 32 data_part_02_02 3931705344 959889 69881856 17061 643.1
> > >
> > > op type count avg. time
> > >
> > > seeks 0 N/A
> > >
> > > reads 0 N/A
> > >
> > > writes 0 N/A
> > >
> > > kaio_reads 260172 0.0013
> > >
> > > kaio_writes 1192 0.0530
> > >
> > > as can be seen 660.4 io/s at 04:00 and 643.1 at 08:00 This looks like
it
> > is
> > > the io/s during some time window - don't know if maximum or average -

> > looks
> > > like maximum..
> > >
> > > So still trying to fully understand what I am seeing
> > >
> > >
> > >
> > >
> >
> >
>
*******************************************************************************

> > > Forum Note: Use "Reply" to post a response in the discussion forum.
> > >
> > >
> >
> > --00504502cc915bd0c304879b33d4
> >
> >
> >
> >
>
*******************************************************************************

> > Forum Note: Use "Reply" to post a response in the discussion forum.
> >
> >
>
> --
> Fernando Nunes
> Portugal
>
> http://informix-technology.blogspot.com
> My email works... but I don't check it frequently...
>
> --0016367f9eee53ab0004879ba4a5
>
>
>
*******************************************************************************

> Forum Note: Use "Reply" to post a response in the discussion forum.
>

Messages In This Thread

• onstat -g iof
  ROBERT CLOW -- Thursday, 27 May 2010, at 8:32 a.m.
  • Re: onstat -g iof
    Fernando Nunes -- Thursday, 27 May 2010, at 8:41 a.m.
    • Re: onstat -g iof
      ROBERT CLOW -- Thursday, 27 May 2010, at 10:30 a.m.
      • Re: onstat -g iof
        Fernando Nunes -- Thursday, 27 May 2010, at 11:10 a.m.
        • Re: onstat -g iof
          ROBERT CLOW -- Thursday, 27 May 2010, at 5:26 p.m.
          • Re: onstat -g iof
            Fernando Nunes -- Thursday, 27 May 2010, at 5:43 p.m.
          • Re: onstat -g iof
            Art Kagel -- Thursday, 27 May 2010, at 6:49 p.m.
            • Re: onstat -g iof
              Fernando Nunes -- Thursday, 27 May 2010, at 7:20 p.m.
              • Re: onstat -g iof
                John Miller iii -- Saturday, 29 May 2010, at 5:47 a.m.
                • Re: onstat -g iof
                  ROBERT CLOW -- Monday, 31 May 2010, at 1:06 a.m.

IDS Forum is maintained by Administrator with WebBBS 5.12.