Re: Control files are updated constantly

From: koert54 <nospam_at_nospam.com>
Date: Wed, 13 Mar 2002 09:59:20 +0100
Message-ID: <3c8f1635$0$7110$4d4efb8e@news.be.uu.net>

Found a filemon output of a system running 10 DB's on AIX with 30 SSA disks in RAID1.
So in the output file you'll see more than one control file all from different databases.
Now what stumbles me is that there are heavy *reads* on the control files (no -
there's no backup running) - who has half an idea of what this is about ?

Suggestions are welcome :-)
This made me remap all database files on the available disks ...

Mon Oct 8 09:10:35 2001
System: AIX

258.278 secs in measured interval
Cpu utilization: 44.9%

Most Active Files

---

  #MBs  #opns   #rds   #wrs  file                 volume:inode
------------------------------------------------------------------------
  95.0      1  12162      0  control01.ctl        /dev/lv18:147457
  85.8      1  10985      0  control01.ctl        /dev/fslv00:16389
  77.4      1   9907      0  control01.ctl        /dev/fslv00:16386
  45.4      1   5812      0  control01.ctl        /dev/lv36:98309
  44.4      1   5681      0  control01.ctl        /dev/lv13:557057
  40.0      4      0    642  archLAND8647.dbf     /dev/lv33:44
  40.0      2    641      0  redo03.log           /dev/lv36:32769
  29.4      7      0    329  redo04.log           /dev/lv36:32771
  29.4      7      0    329  redo14.log           /dev/lv25:1228804
  25.0      1   6401      0  control01.ctl        /dev/fslv20:8197
  21.1      2   5390      0  control01.ctl        /dev/lv17:229377
  20.4    119   5235      0  unix                 /dev/hd2:8492
   5.0      2   1280      0  control01.ctl        /dev/lv18:16388
   4.0    234    468      0  lx1boot.nlb          /dev/lv14:82110
   3.8      1    486      0  control01.ctl        /dev/fslv05:1687553
   3.5      3    886      0  DATA01.DBF           /dev/lv18:16397
   3.0    129    764      0  hosts                /dev/hd4:4225
   2.8      5    715      0  land_indx01.dbf      /dev/lv17:229386

"Connor McDonald" <connor_mcdonald_at_yahoo.com> wrote in message news:3C8E6850.7419_at_yahoo.com...
> koert54 wrote:
> >
> > I've noticed this too - I used to think that there isn't heavy I/O on
the
> > controlfiles but checking with filemon on AIX it turned out that
> > controlfiles
> > DO get hit with massive I/O !!! I still don't understand what the I/O is
> > about, except for some checkpointing, but still ... maybe someone can
> > clarify this ???
> >
> > Knowing this - I never use more than 2 controlfiles (throw away that
third
> > one create by oracle ... don't need more I/O) and I never put them on
the
> > same disks
> > as system and redologs ...
> >
> > "Orlando Amador" <oamador_at_psgapr.jnj.com> wrote in message
> > news:6e75aa9.0203121033.1401091f_at_posting.google.com...
> > > Using a file monitoring tool from www.sysinternals.com it looks like
> > > the database control files are being updated every 2 o 3 seconds. The
> > > database is fairly small and other files like redo, or data files are
> > > not being updated that frequently.
> > >
> > > Is it normal for oracle.exe to write every few seconds to control
> > > files?
> > > Is this adjustable?
> > >
> > > Saludos,
> > >
> > > Orlando M. Amador
> > > Ortho Biologics LLC
> > > Sr. System Analyst
> > > oamador_at_psgapr.jnj.com

>

> The ckpt process updates the control file every 3 seconds with
> checkpoint pointer information.

>

> hth
> connor
> --
> ==============================
> Connor McDonald

>

> http://www.oracledba.co.uk

>
> "Some days you're the pigeon, some days you're the statue..."
Received on Wed Mar 13 2002 - 02:59:20 CST