Re: [K12OSN] performance comparison

• Messages sorted by: [ date ][ subject ][ author ]
• Next message: Steve Wright "Re: [K12OSN] performance comparison"
• Previous message: pedro noticioso "Re: [K12OSN] Netscape vs. Mozilla thickness?"

--=-LlYvW4qrFpFiyGF+jo3E
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

Op za 01-03-2003, om 15:09 schreef BMan:

:)

Yep the ide's are 133,

The hardware scsi-raid is an addoncard on top of dual scsi320
controller.
I set that also up without making use of the hardware raid setup ie just
plain 3 array's and then with the linux software raid with the disks
over two controllers (The RAID card forces the disks into one
controller)

The performance was identical.

Actually I'm more surprised about the fact the the individual scsi disks
give 36 MB/s (IBM says 56 and 76 top), but three in a stripe hardly
doubles the performance i.e. 56MB/s. cache was setup as write-back.

It shouldn't be the controller with 320 throughput...
hmmm, strange...

Still I prefere the scsi solution because of its non performance
degrading qualities on sustained random read/writes by multiple users at
the same time.

=20

> Psssst, don't tell anyone.  ;)  Are the drives on the promise controller=20
> ATA-100 or 133?  A good test is to compare the controller based RAID=20
> Solution vs. built in software RAID.  You may find the extra expense for=20
> the controllers can be avoided too.  :)
>=20
> FYI,  ALL Red Hat servers run software RAID without any issues. I also=20
> run software RAID on all solutions I setup.   :)
>=20
> Regards,
> BMan
>=20
> Immanuel Derks wrote:
>=20
> >Hi,
> >
> >I'm just a bit curious on some performance issue concerning scsi-raid
> >setup versus ide software-raid:
> >
> >I run a stripe from a IBM ServerRAID 5i with 3 36GB/15000RPM disks
> >(RAID00). The disks where all set to write-back i.s.o write-through.
> > and hdparm (-tT) gives me:
> >
> >hdparm -tT /dev/sda:
> >Timing buffer-cache reads: 128 MB in 0.28 seconds=3D461.55 MB/s
> >Timing buffered disk reads: 64 MB in 1.12 secons =3D 56.98 MB/sec
> >
> >
> >compared to a software raid setup on a promise controller with two
> >WesternDigitals 80Gig 8MB cache ide drives:
> >
> >hdparm -tT /dev/md0
> >/dev/md0:
> >Timing buffer-cache reads: 128MB in 0.36 seconds =3D 352.37 MB/sec
> >Timing buffered disk reads: 64 MB in 1.25 seconds =3D 51.20 MB/sec
> >
> >Isn't that curious, there is hardly any difference between mega
> >expensive IBM setup and cheap of the shelve components....
> >
> >Immanuel
> >
> >=20
> >
> >
> > =20
> >
>=20
>=20
>=20
>=20


--=-LlYvW4qrFpFiyGF+jo3E
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: Dit berichtdeel is digitaal gesigneerd

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQA+YRXO5NWkbih6y88RAnE/AJ9e8BeWCfLXN410V7DGubAq1hnlAQCggFCZ
gI7DteyJmX+XX4LJhERKcN4=
=P+eR
-----END PGP SIGNATURE-----

--=-LlYvW4qrFpFiyGF+jo3E--



_______________________________________________
K12OSN mailing list
K12OSN@redhat.com
https://listman.redhat.com/mailman/listinfo/k12osn
For more info see 

• Next message: Steve Wright "Re: [K12OSN] performance comparison"
• Previous message: pedro noticioso "Re: [K12OSN] Netscape vs. Mozilla thickness?"