SAN disks
We have some LUNs set up on a Dell (EMC/Clarion) SAN for our main databases. EMC is telling us that we will get better performance if we set up our db spaces as filesystems instead of raw disk. Has anyone actually tested I/O performance for IDS for raw disks on an EMC (or other vendor) vs cooked files in file systems? Bob Roussey Unix / Informix Administration Spirit Airlines Robert.Roussey@SpiritAir.com
Robert Roussey\\\\(IT\\\\) said: > We have some LUNs set up on a Dell (EMC/Clarion) SAN for our main > databases. EMC is telling us that we will get better performance if we > set up our db spaces as filesystems instead of raw disk. Just tell them to fuck off and leave you alone. -- Bye now, Obnoxio "I'm astonished anyone pays real money for this crap." -- Cosmo -- This message has been scanned for viruses and dangerous content by OpenProtect(http://www.openprotect.com), and is believed to be clean.
I did but long ago. Can't hurt to benchmark this again. If you write your own test app, as I did once (no I don't know what happened to the source), make sure to open the COOKED devices or filesystem chunk files with O_SYNC which is how IDS does it to insure data integrity. Without O_SYNC the 'apparent' performance will be great but the data will not have been flushed to the physical drives until long after the writes return. Art S. Kagel ----- Original Message ----- From: Robert Roussey <ids@iiug.org> At: 4/16 11:54:56 We have some LUNs set up on a Dell (EMC/Clarion) SAN for our main databases. EMC is telling us that we will get better performance if we set up our db spaces as filesystems instead of raw disk. Has anyone actually tested I/O performance for IDS for raw disks on an EMC (or other vendor) vs cooked files in file systems? Bob Roussey Unix / Informix Administration Spirit Airlines Robert.Roussey@SpiritAir.com ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum.
I suggest you ask them to prove it. I use Clariion CX-700 and they've never told me so huge absurd. Celso Coimbra -----Mensagem original----- De: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org]Em nome de Obnoxio The Clown Enviada em: segunda-feira, 16 de abril de 2007 13:05 Para: ids@iiug.org Assunto: Re: SAN disks [8914] Robert Roussey\\\\(IT\\\\) said: > We have some LUNs set up on a Dell (EMC/Clarion) SAN for our main > databases. EMC is telling us that we will get better performance if we > set up our db spaces as filesystems instead of raw disk. Just tell them to fuck off and leave you alone. -- Bye now, Obnoxio "I'm astonished anyone pays real money for this crap." -- Cosmo -- This message has been scanned for viruses and dangerous content by OpenProtect(http://www.openprotect.com), and is believed to be clean. ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum.
On 4/16/07, ART KAGEL, BLOOMBERG/ 731 LEXIN <kagel@bloomberg.net> wrote: > I did but long ago. Can't hurt to benchmark this again. If you write your own > test app, as I did once (no I don't know what happened to the source), make > sure > to open the COOKED devices or filesystem chunk files with O_SYNC which is how > IDS does it to insure data integrity. Without O_SYNC the 'apparent' > performance > will be great but the data will not have been flushed to the physical drives > until long after the writes return. > > Art S. Kagel > > ----- Original Message ----- > From: Robert Roussey <ids@iiug.org> > At: 4/16 11:54:56 > > We have some LUNs set up on a Dell (EMC/Clarion) SAN for our main > databases. EMC is telling us that we will get better performance if we > set up our db spaces as filesystems instead of raw disk. > Has anyone actually tested I/O performance for IDS for raw disks on an > EMC (or other vendor) vs cooked files in file systems? Also, he says, bottom posting as in the FAQ, consider the Direct I/O options (O_DIRECT). The latest versions of IDS use this - or will use this once released - on platforms where it is available. It may never appear in 9.40; it might or might not be in 10.00; it most probably will be in Cheetah. -- Jonathan Leffler #include <disclaimer.h> Email: jleffler@earthlink.net, jleffler@us.ibm.com Guardian of DBD::Informix v2007.0226 -- http://dbi.perl.org/ NB: Please do not use this email for correspondence. I don't necessarily read it every week, even.