"Art S. Kagel" <kagel@bloomberg.net> wrote in message news:<pan.2004.04.02.08.20.02.153746.1
445@bloomberg.net>...
> On Thu, 01 Apr 2004 03:57:48 -0500, Andy Kent wrote:
>
> I mostly agree with Andy, put everything on the single RAID10 array.  If t
he
> server were to be heavily loaded with update activity then perhaps you cou
ld
> make a case for reserving one pair of drives for a mirror (RAID1) to conta
in
> just the ROOTDB and LOGSDB.  But if the load will not be high, I would not
> bother.
>
> Art S. Kagel
>
> PS: Remember to stripe the 5 mirrored pairs not the other way around.

Hi Art,
You lost me here, what exactly did you mean by "PS: Remember to stripe
the 5 mirrored pairs not the other way around."?

[vbcol=seagreen]
> 

[ Post a follow-up to this message ]

You want to mirror then stripe not the other way round.  If you do the
latter then a single disk failure on both stripes will take out the
entire disk.  If you mirror first the system has more redundancy.

If in doubt the draw it:-))

DNix wrote:[vbcol=seagreen]
>
> "Art S. Kagel" <kagel@bloomberg.net> wrote in message news:<pan.2004.04.02
.08.20.02.153746.1445@bloomberg.net>... 
>
> Hi Art,
> You lost me here, what exactly did you mean by "PS: Remember to stripe
> the 5 mirrored pairs not the other way around."?
> 

--
Paul Watson             #
Oninit Ltd              # Growing old is mandatory
Tel: +44 1436 672201    # Growing up is optional
Fax: +44 1436 678693    #
Mob: +44 7818 003457    #
www.oninit.com          #

[ Post a follow-up to this message ]

On Mon, 19 Apr 2004 09:30:39 -0400, DNix wrote:

Paul's got it.  To build RAID10 (versus RAID01) first you create N mirrored
pairs then you stripe the mirror devices.  This gives you greater safety and
improved recovery time over RAID01.  Performance is identical with either
RAID10 or RAID01.

Art S. Kagel
[vbcol=seagreen]
> "Art S. Kagel" <kagel@bloomberg.net> wrote in message
> news:<pan.2004.04.02.08.20.02.153746.1445@bloomberg.net>... 
>
> Hi Art,
> You lost me here, what exactly did you mean by "PS: Remember to stripe the
 5
> mirrored pairs not the other way around."?
>
>
> 

[ Post a follow-up to this message ]

You might want to check the documentation on the raid mangler you
are using, e.g under Sun RM6 RAID10 is referred to as RAID0+1, under
DiskSuite AFAIR you can't RAID10 you must RAID01 etc

Paul Watson wrote:
>
> You want to mirror then stripe not the other way round.  If you do the
> latter then a single disk failure on both stripes will take out the
> entire disk.  If you mirror first the system has more redundancy.
>
> If in doubt the draw it:-))
>
> DNix wrote: 
>
> --
> Paul Watson             #
> Oninit Ltd              # Growing old is mandatory
> Tel: +44 1436 672201    # Growing up is optional
> Fax: +44 1436 678693    #
> Mob: +44 7818 003457    #
> www.oninit.com          #

--
Paul Watson             #
Oninit Ltd              # Growing old is mandatory
Tel: +44 1436 672201    # Growing up is optional
Fax: +44 1436 678693    #
Mob: +44 7818 003457    #
www.oninit.com          #

[ Post a follow-up to this message ]

In article <4084073B.75F6DFFD@oninit.com>,
Paul Watson <paul@oninit.com> wrote:

> You might want to check the documentation on the raid mangler you
> are using, e.g under Sun RM6 RAID10 is referred to as RAID0+1, under
> DiskSuite AFAIR you can't RAID10 you must RAID01 etc

Supposedly DS a.k.a SVM recognizes your RAID01 setup and changes it into
RAID10 behind the scenes -- even though you define the metadevices as
if they were RAID01.  I've never attempted to verify this, mind.

Karl

[ Post a follow-up to this message ]

That's rather disturbing given the earlier comments. I think it may
well be DiskSuite.

Paul Watson <paul@oninit.com> wrote in message news:<4084073B.75F6DFFD@oninit.com>...
> You might want to check the documentation on the raid mangler you
> are using, e.g under Sun RM6 RAID10 is referred to as RAID0+1, under
> DiskSuite AFAIR you can't RAID10 you must RAID01 etc

[ Post a follow-up to this message ]

Maybe I'm a little off track here, but how about placing them on the server'
s internal disks ?    Just a thought ....... didn't follow the entire thread
 ...







-----Original Message-----

From: owner-informix-list@iiug.org [mailto:owner-informix-list@iiug.org]
 On Behalf Of DNix

Sent: Monday, April 19, 2004 3:31 PM

To: informix-list@iiug.org

Subject: Re: Where to put logs in a RAUD 1+0 system



"Art S. Kagel" <kagel@bloomberg.net> wrote in message news:<pan.2004.04.02.0
8.20.02.153746.1445@bloomberg.net>...

> On Thu, 01 Apr 2004 03:57:48 -0500, Andy Kent wrote:

>

> I mostly agree with Andy, put everything on the single RAID10 array.  If the[/vbco
l]
[vbcol=seagreen]
> server were to be heavily loaded with update activity then perhaps you could[/vbco
l]
[vbcol=seagreen]
> make a case for reserving one pair of drives for a mirror (RAID1) to contain[/vbco
l]
[vbcol=seagreen]
> just the ROOTDB and LOGSDB.  But if the load will not be high, I would not

> bother.

>

> Art S. Kagel

>

> PS: Remember to stripe the 5 mirrored pairs not the other way around.



Hi Art,

You lost me here, what exactly did you mean by "PS: Remember to stripe

the 5 mirrored pairs not the other way around."?





>
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 
[vbcol=seagreen] 




________________________________

<< ella for Spam Control >> has removed 325 Spam messages and set aside 270 
Newsletters for me
You can use it too - and it's FREE!  www.ellaforspam.com

sending to informix-list

[ Post a follow-up to this message ]

I found a paper written for an RDBMS whose name escapes me which
rather answers my question:

<snip>
In general it is easiest and most efficient to stripe the logs across
all the disks just like the data files. Sometimes
people worry that placing the log files on the same disks as the data
files will cause interference between data accesses
and log writes. This is because the disk head may have to move to a
new position when the log is written. As we
discussed in the previous section, for relatively small seeks the
rotational latency of the IO will dominate the seek
time. So, if the log file is placed along with the other frequently
accessed data on the outside half of the disk,
interference will not be a significant problem. Striping across too
few disks is a bigger problem in practice.
</snip>

[ Post a follow-up to this message ]

> an RDBMS whose name escapes me which

Oracle - the SAME (Stripe And Mirror Everything) paper.

That's 2 for 2 in the naming department :-)

[ Post a follow-up to this message ]