Re: Async_io set up HP-UX for 10g Instances(these are apps instances)

  • From: "LS Cheng" <exriscer@xxxxxxxxx>
  • To: "edwin devadanam" <edwin_kodamala@xxxxxxxxx>
  • Date: Mon, 6 Aug 2007 17:57:58 +0200

Simulate, not AIO!


Thanks

--
LSC


On 8/6/07, edwin devadanam <edwin_kodamala@xxxxxxxxx> wrote:
>
> Cheng,
>
> metalink note: 47328.1  says:
> If you set DISK_ASYNCH_IO to false, then you should also set
> DBWR_IO_SLAVES to a value
> other than its default of zero in order to simulate asynchronous I/O.
>
>
>
> Regards,
> Edwin.K
>
> *LS Cheng <exriscer@xxxxxxxxx>* wrote:
>
> Dude we all have told you you cannot use AIO with Filesystems!
>
> If you really really want to use AIO with Filesystems then buy Veritas
> Quick I/O. That does support AIO with Filesystem but using Cluster
> Filesystems.
>
> Thanks
>
> --
> LSC
>
>
> On 8/6/07, edwin devadanam <edwin_kodamala@xxxxxxxxx> wrote:
> >
> > Hi All,
> >
> > Please add ur valuable comments on the below mentioned points.
> >
> > This is production database:
> >
> > SQL> show sga
> > Total System Global Area 3215950016 bytes
> > Fixed Size                   741568 bytes
> > Variable Size            1056964608 bytes
> > Database Buffers         2147483648 bytes
> > Redo Buffers               10760192 bytes
> >
> > 1) As discussed we can't use disk_asynch_io=true in HP-UX(its using jfs
> > not raw filesystem)
> > 2) Is HP-UX(JFS) doesnt suport asynchronous io?(operating system level)
> > 3) dbwr_io_slaves=x or db_writer_processes=y which one does asynchronous
> > i/o?(refer Note:47328.1)
> > 4)Keeping db_writer_processes=y only will i not achieve asynchornous
> > i/o?
> >
> > Thanks In Advance,
> > Edwin.K
> >
> > *Roby Sherman <rxsherm@xxxxxxxxxxxxx>* wrote:
> >
> > >
> > > - disk_asynch_io is used to enable the oracle kernel to make use of
> > > asynchronous io, IF the OS supports it.
> >
> > And you really DO want to figure out if your OS supports it.
> > Otherwise if you turn it on and its not supported, your database is
> > essentially making two I/O calls for every one... One to aio_write
> > (which fails) and then an immediate call to pwrite after. I can't
> > tell you how many environments I come into where disk_asynch_io is
> > set to true and it isn't supported.
> >
> > --Roby
> > --
> > //www.freelists.org/webpage/oracle-l
> >
> >
> >
> >  ------------------------------
> > Be a better Globetrotter. Get better travel answers
> > <http://us.rd.yahoo.com/evt=48254/*http://answers.yahoo.com/dir/_ylc=X3oDMTI5MGx2aThyBF9TAzIxMTU1MDAzNTIEX3MDMzk2NTQ1MTAzBHNlYwNCQUJwaWxsYXJfTklfMzYwBHNsawNQcm9kdWN0X3F1ZXN0aW9uX3BhZ2U-?link=list&sid=396545469>from
> > someone who knows.
> > Yahoo! Answers - Check it out.
> >
>
>
> ------------------------------
> Need a vacation? Get great deals to amazing places
> <http://us.rd.yahoo.com/evt=48256/*http://travel.yahoo.com/;_ylc=X3oDMTFhN2hucjlpBF9TAzk3NDA3NTg5BHBvcwM1BHNlYwNncm91cHMEc2xrA2VtYWlsLW5jbQ-->on
> Yahoo! Travel.
>
>

Other related posts: