Re: NT vs Linux as web server
I clicked on the link a couple of minutes ago. It still hasn't come up! (ok, so it's probably the network in between, but I thought that was kinda ironic in the Alanis Morissette sense of the word) Sorry for the pointless posting: I'm supposed to be revising! Rich Peter S Galbraith wrote: My IT manager just EMailed me this article (CC'ed to a bunch of Directors, of course): http://www.mindcraft.com/whitepapers/nts4rhlinux.html Microsoft Windows NT Server 4.0 is 2.5 times faster than Linux as a File Server and 3.7 times faster as a Web Server. I'm sure I could dig up opposing reviews. Anyone know of any? Peter -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: NT vs Linux as web server
On Wed, Apr 14, 1999 at 10:45:01AM -0400, Peter S Galbraith wrote: My IT manager just EMailed me this article (CC'ed to a bunch of Directors, of course): http://www.mindcraft.com/whitepapers/nts4rhlinux.html Microsoft Windows NT Server 4.0 is 2.5 times faster than Linux as a File Server and 3.7 times faster as a Web Server. I'm sure I could dig up opposing reviews. Anyone know of any? Well, start with zdnet, who did reviews with the exact same benchmarks and came to almost the opposite conclusions. Also, try lwn.net, which is compiling a list of grevious errors in this study. -- Ian Peters I never let schooling interfere with my education. [EMAIL PROTECTED] -- Mark Twain
Re: NT vs Linux as web server
Peter S Galbraith ([EMAIL PROTECTED]) said: My IT manager just EMailed me this article (CC'ed to a bunch of Directors, of course): http://www.mindcraft.com/whitepapers/nts4rhlinux.html Microsoft Windows NT Server 4.0 is 2.5 times faster than Linux as a File Server and 3.7 times faster as a Web Server. I'm sure I could dig up opposing reviews. Anyone know of any? Linux Weekly News (www.lwn.com) is formulating a reply about the inconsistencies/inaccuracies of those tests (I believe the samba server was somewhat crippled among other things), not to mention that they were sponsored BY Microsoft. Check out the response on Slashdot (www.slashdot.org) for other problems. As my Probability and Statistics professor says you can make statistics say whatever you want, but it's not always accurate .adam -- Adam Lazur - Computer Engineering Undergrad - Lehigh University icq# 3354423 - http://www.lehigh.edu/~ajl4 Besides, I think Slackware sounds better than 'Microsoft,' don't you? -Patrick Volkerding
Re: NT vs Linux as web server
On Wed, 14 Apr 1999, Peter S Galbraith wrote: My IT manager just EMailed me this article (CC'ed to a bunch of Directors, of course): http://www.mindcraft.com/whitepapers/nts4rhlinux.html Microsoft Windows NT Server 4.0 is 2.5 times faster than Linux as a File Server and 3.7 times faster as a Web Server. I'm sure I could dig up opposing reviews. Anyone know of any? Peter- 1) The white-paper was commissioned by MS. It's right there in the paper. That's the most telling fact in the whole paper. 2) http://lwn.net/1999/features/MindCraft.phtml has a list of critiques of the proposal, including the suggestion that they deliberately used a kernel (2.2.2) with known networking problems. They also have a list of links with research you can use to counter theirs, from several respected and independent news sources. 3) http://slashdot.org/article.pl?sid=99/04/14/0042212 is /.'s thread on this- lots of interesting observations and criticisms. Make sure you set Highest Scores First- otherwise you will have to search forever to find the pertinent ones. Good luck- I'd strongly suggest sending out at least the lwn.net link to counter the FUD. -Luis ### They call the faithful to their knees to hear the softly spoken magic spell: There's no place like home... There's no place like home... There's no place like home. -Pink Floyd, Dark Side of the Moon -Dorothy, The Wizard of Oz ###
Re: NT vs Linux as web server
-BEGIN PGP SIGNED MESSAGE- On Wed, 14 Apr 1999, Peter S Galbraith wrote: My IT manager just EMailed me this article (CC'ed to a bunch of Directors, of course): http://www.mindcraft.com/whitepapers/nts4rhlinux.html Microsoft Windows NT Server 4.0 is 2.5 times faster than Linux as a File Server and 3.7 times faster as a Web Server. I'm sure I could dig up opposing reviews. Anyone know of any? Actually, you can find several opposing views directly in the white paper. First of all, the test was sponsored by MS. Try finding an independant test and check the results. ZDnet did one a while back with very different results. Here are a couple of links to check out: http://www.zdnet.com/sr/stories/issue/0,4537,2196106,00.html http://www.zdnet.com/sr/stories/issue/0,4537,396321,00.html Note that these links only really talk about file serving, not web serving. However, they do take some credibility away from the mindcraft survey. There is also a response to the study over at Linux Weekly News: http://lwn.net/1999/features/MindCraft.phtml It appears as though Mindcraft spent quite a bit of time tuning NT, and very little time tuning Linux. So, I suppose you should start at the links I've given here. You also might want to talk to some people who use Linux every day for high volume web serving. Rob Malda at Slashdot.org would be worth talking to. His site gets a huge number of hits every day, and really performs quite well considering the amount of dynamic content. noah PGP public key available at http://lynx.dac.neu.edu/home/httpd/n/nmeyerha/mail.html or by 'finger -l [EMAIL PROTECTED]' This message was composed in a 100% Microsoft free environment. -BEGIN PGP SIGNATURE- Version: 2.6.2 iQCVAwUBNxTEm4dCcpBjGWoFAQF4hQP+LvVsj/m8bqr80UJnb5AyGjwq8adLnF7Z 3Y8VSAxq5dJXq2MykdrH9tF/WwO0Pt8jlYvx4uzU1aNSyXLgdIXJ5g48JrlofG+p /Kyiv8H9xlTUUkSyPCGrbnlJs1XSGV0GidOgQk1BuyLw3Na1CERlJfl5U6NRl9Al uwewmcSWWOk= =UfiY -END PGP SIGNATURE-
Re: NT vs Linux as web server
note the following about 4/5 of the way through Mindcraft, Inc. conducted the performance tests described in this report between March 10 and March 13, 1999. Microsoft Corporation sponsored the testing reported herein. -Michael On Wed, 14 Apr 1999, Peter S Galbraith wrote: My IT manager just EMailed me this article (CC'ed to a bunch of Directors, of course): http://www.mindcraft.com/whitepapers/nts4rhlinux.html Microsoft Windows NT Server 4.0 is 2.5 times faster than Linux as a File Server and 3.7 times faster as a Web Server. I'm sure I could dig up opposing reviews. Anyone know of any? Michael Stenner Office Phone: 919-660-2513 Duke University, Dept. of Physics [EMAIL PROTECTED] Box 90305, Durham N.C. 27708-0305
Re: NT vs Linux as web server
There have been a lot of discussion on this benchmark on slashdot (http://www.slashdot.org). I had time to take a galnce and it seems that the benchmark is biased. It seems they have done a very good tunning of the NT box and a poor one for the linux box. As a small exemple they have used a server with 4GB of RAM. NT could handle it, but they claim taht linux (kernel 2.2) did recognize only 1 GB. I may be confused but doesn't the new kernel support at least 2GB (I am sure I have seen some VA research workstations with 2GB). What is the maximum linux kernel can handle? Paulo. -- Paulo José da Silva e Silva [EMAIL PROTECTED] Ph.D. Student in Applied Math. University of São Paulo - Brazil http://www.ime.usp.br/~rsilva May the code be with you :-)
Re: NT vs Linux as web server
I have just read the lwn comments. They have pointed out that the NT server was setted to use only 1GB of memory, so my last example of biased tunning doens't apply. Sorry for my error :-). Any way I would be glad to know which is the maximum amount of RAM kernel 2.2 can handle. Thank you all, Paulo -- Paulo José da Silva e Silva [EMAIL PROTECTED] Ph.D. Student in Applied Math. University of São Paulo - Brazil http://www.ime.usp.br/~rsilva May the code be with you :-)
Re: NT vs Linux as web server
The March 22 issue of Smart Reseller (www.smartreseller.com) compared NT and Linux running Samba and it had Linux/Samba way ahead. So I was very surprized to see the test by Mindcraft. Try the following: www.zdnet.com/sr/stories/infopack/0,5483,387506,00.html There are two links on that page -- one for Samba, one for Apache. In both articles, NT fails in the 10 to 12 user area. Good luck! -- Gregory Wood Farsight Computer 1219 W University Blvd Odessa TX 79764 Voice: 1-915-335-0879 Member: CT Pioneers Luis Villa wrote: On Wed, 14 Apr 1999, Peter S Galbraith wrote: My IT manager just EMailed me this article (CC'ed to a bunch of Directors, of course): http://www.mindcraft.com/whitepapers/nts4rhlinux.html Microsoft Windows NT Server 4.0 is 2.5 times faster than Linux as a File Server and 3.7 times faster as a Web Server. I'm sure I could dig up opposing reviews. Anyone know of any? Peter- 1) The white-paper was commissioned by MS. It's right there in the paper. That's the most telling fact in the whole paper. 2) http://lwn.net/1999/features/MindCraft.phtml has a list of critiques of the proposal, including the suggestion that they deliberately used a kernel (2.2.2) with known networking problems. They also have a list of links with research you can use to counter theirs, from several respected and independent news sources. 3) http://slashdot.org/article.pl?sid=99/04/14/0042212 is /.'s thread on this- lots of interesting observations and criticisms. Make sure you set Highest Scores First- otherwise you will have to search forever to find the pertinent ones. Good luck- I'd strongly suggest sending out at least the lwn.net link to counter the FUD. -Luis ### They call the faithful to their knees to hear the softly spoken magic spell: There's no place like home... There's no place like home... There's no place like home. -Pink Floyd, Dark Side of the Moon -Dorothy, The Wizard of Oz ### -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null -- Gregory Wood Farsight Computer 1219 W University Blvd Odessa TX 79764 Voice: 1-915-335-0879 Member: CT Pioneers
Re: NT vs Linux as web server
Spring 1999 Issue of linux magazine, page 42: LINUX OUTPERFORMED WINDOWS by as much as 250% for 12 or more client systems. (emphasis theirs, this is regarding SAMBA) If I may say so, both sides seem to be generating a lot of FUD on this. In my own (unscientific) studies, Linux has outperformed NT, but only because Linux is operating without a processor-intensive GUI, and without other unnecessary (for a file server, anyway) support services (which are darn near impossible to remove on an NT Server). Generally, the most important things to consider on these X is faster than Y comparisons is to check the science behind the comparison. If Windows NT is faster than Linux on a two-machine network, how does that matter to you on your 100 machine LAN? If the article is hesitant to describe the methodology behind their study, and if it sounds too much like laboratory conditions, than the study is bogus. On Wed, 14 Apr 1999, Peter S Galbraith wrote: My IT manager just EMailed me this article (CC'ed to a bunch of Directors, of course): http://www.mindcraft.com/whitepapers/nts4rhlinux.html Microsoft Windows NT Server 4.0 is 2.5 times faster than Linux as a File Server and 3.7 times faster as a Web Server. I'm sure I could dig up opposing reviews. Anyone know of any? Peter -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: NT vs Linux as web server
Itf your looking for articles look at slashdot.org's achrive. But if I'm correct(I'd head to double check ) I belive the fine print say Micosoft payed for it. Also the configuration I believed was such that they would either cripple Linux or not optimize it liek they fine tuned NT. I could be wrong though... Philip Thiem(my backsspace is current broken is please ec[Dxcuse the typoes[D[Ds )
Re: NT vs Linux as web server
Adam Lazur ([EMAIL PROTECTED]) said: ---SNIP--- Linux Weekly News (www.lwn.com) is formulating a reply about the ^ doh, make that .net -- Adam Lazur - Computer Engineering Undergrad - Lehigh University icq# 3354423 - http://www.lehigh.edu/~ajl4 Besides, I think Slackware sounds better than 'Microsoft,' don't you? -Patrick Volkerding
RE: NT vs Linux as web server
I clicked on the link a couple of minutes ago. It still hasn't come up! (ok, so it's probably the network in between, but I thought that was kinda ironic in the Alanis Morissette sense of the word) Sorry for the pointless posting: I'm supposed to be revising! Rich Came up fast for me. Also read it and saw that Microsoft sponsored the test... You want comments, look at slashdot.org - there's almost 600 of them! Peter S Galbraith wrote: My IT manager just EMailed me this article (CC'ed to a bunch of Directors, of course): http://www.mindcraft.com/whitepapers/nts4rhlinux.html Microsoft Windows NT Server 4.0 is 2.5 times faster than Linux as a File Server and 3.7 times faster as a Web Server. I'm sure I could dig up opposing reviews. Anyone know of any? Peter -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: NT y LINUX
El miércoles 07 de abril de 1999 a la(s) 11:23:30 +0100, Jose Marin contaba: comment No mandes HTML a la lista, please, que queda feo. /comment El mutt soporta MIME y al ver el attachment me llamó al lynx automáticamente. Aunque no deja de ser una pesadez. Podrias explicar un poco mejor lo que dices aqui? Si leo bien, el bootloader del LILO lo pones en /dev/hda2. Por lo cual, al arrancar, el bootloader que se cargaria es el que sigue estando en el MBR, el de WinNT (suponiendo que NT fue instalado antes que Linux). O donde me equivoco...? Con fdisk, se activa la partición de Linux y así al arrancar, se va directamente a Lilo, desde que el que llamamos a eneté si hace falta. Jose L. Marín [EMAIL PROTECTED] Dept of Maths [EMAIL PROTECTED] -- Just do it. David Serrano [EMAIL PROTECTED] Linux Registered User no. 87069 http://come.to/Hue-Bond.world In love with TuX. Linux 2.2.5 PGP Public key at http://www.ctv.es/USERS/fserrano/pgp_pubkey.asc
Re: NT y LINUX
Han Solo wrote: On Mon, Mar 29, 1999 at 03:58:17PM +0200, Ramiro Alba wrote: > > Tenemos Windows NT 4.0 instalado en una particin del primer disco y en > otra(s) > particiones del mismo disco instalamos Debian y onfiguramos Lilo para > que arranque de los 2 sistemas. El arranque de Linux ningn problema > pero el de NT comienza bien hasta que aparece la pantallita azul y > despues de unos 10 segundos falla estrepitosamente. Esto no ocurre si el > disrectorio root de Linux esta en una particion de otro discos. Si no es > el caso, me he visto obligado a poner el Lilo en disket para el > arranque dual. Hasta donde he podido me mirado la documentacin de Lilo > a fondo, pero no he dado con la causa. Alguien sabe que demonios pasa? > Yo lo planteara de otra manera, que es la que a mi me ha funcionado: instalas lilo en la particin que va contener a linux, siempre por debajo del cilindro 1024 y en disco master. Luego copias el sector de arranque de linux en un fichero, con dd if=/dev/hda2 of=/bootsect.lnx bs=521 count=1 (suponiendo que linux est en hda2). Lo siguiente es copiar el archivo /bootsect.lnx a c: Si c: es una particin ntfs, tendrs que copiarlo primero en un disco. Entonces editas el archivo boot.ini, que es (hablo de memoria) hidden,read-only,system. Antes de editarlo tendrs que cambiarle los atributos, pero luego acurdate de dejarlos como estaban. Como deca, editas el fichero y aades la lnea c:\bootsect.lnx="Linux" Con esto, sers capaz de arrancar linux desde el cargador de NT. Funciona perfectamente (doy fe); el nico inconveniente es que cada vez que retocas el lilo, tienes que copiar el sector de arranque de nuevo. Todo esto viene mejor explicado en el howto Linux+NT+loader (creo que se llamaba as) y en el nmero 5 de Linux Actual. -- Un Saludo Han Solo The Rebel Alliance Conecto, luego existo. Desconecto, luego insisto. Soy usuario de infobirria+ P.D. La firma no es ma, sino de uno que trabajaba, precisamente, en M$. Vivir para ver. -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null Esa solucin es perfectamente vlida, pero creo que pierdes la funcionalidad del cargador LILO. Copia el contenido de lo siguiente y ajustalo a tus necesidades, lo importante es que la particin de instalacin del sector de arranque de LILO sea la que corresponde al File System root de linux, de sta forma puedes mantener los 2 cargadores (LILO y NT Loader) en 2 niveles y desde LILO o bien lanzar Debian o el NT Loader. boot=/dev/hda2 compact prompt install=/boot/boot.b map=/boot/map vga=normal # Imagen lanzada por defecto other=/dev/hda1 label="Windows NT 4.0" table=/dev/hda image=/vmlinuz label="Debian 2.0 Hamm" root=/dev/hda2 read-only Lo anterior me funciona perfectamente sobre un portatil, disco de 4 Gb y particiones de NT (NTFS) de 1.5 Gb y 512 Mb. Espero que les sea de ayuda. begin:vcard fn:Manuel Batista Dominguez n:Batista Dominguez;Manuel email;internet:[EMAIL PROTECTED] tel;work:928 29 64 50 x-mozilla-cpt:;0 x-mozilla-html:FALSE version:2.1 end:vcard
Re: NT y LINUX
On Wed, 7 Apr 1999, Manuel Batista Dominguez wrote: nbsp; Esa solucioacute;n es perfectamente vaacute;lida, pero creo que pierdes la funcionalidad del Bcargador LILO./B BRBnbsp;/B Copia el contenido de lo siguiente y ajustalo a tus necesidades, lo importante es que la particioacute;n de instalacioacute;n del sector de arranque de LILO sea la que corresponde al File System root de linux, de eacute;sta forma puedes mantener los 2 cargadores (LILO y NT Loader) en 2 niveles y desde LILO o bien lanzarnbsp; Debian o el NT Loader. Pboot=/dev/hda2 BRcompact BRprompt BRinstall=/boot/boot.b BRmap=/boot/map BRvga=normal BR# Imagen lanzada por defecto BRother=/dev/hda1 BRnbsp;nbsp;nbsp;nbsp; label=Windows NT 4.0 BRnbsp;nbsp;nbsp;nbsp; table=/dev/hda BRimage=/vmlinuz BRnbsp;nbsp;nbsp;nbsp; label=Debian 2.0 Hamm BRnbsp;nbsp;nbsp;nbsp; root=/dev/hda2 BRnbsp;nbsp;nbsp;nbsp; read-only comment No mandes HTML a la lista, please, que queda feo. /comment Podrias explicar un poco mejor lo que dices aqui? Si leo bien, el bootloader del LILO lo pones en /dev/hda2. Por lo cual, al arrancar, el bootloader que se cargaria es el que sigue estando en el MBR, el de WinNT (suponiendo que NT fue instalado antes que Linux). O donde me equivoco...? Y si es asi, que ventaja hay en tener una seccion para WinNT en lilo.conf? Supongo que lo mejor seria poner LILO en el MBR (i.e., boot=/dev/hda), y tenerlo asi de master bootloader. Qué creeis? Pero, en ese caso, sabe alguien como guardar el MBR original (el bootloader de NT), por si interesa dejarlo como estaba en un futuro? JL = Jose L. Marín [EMAIL PROTECTED] Dept of Maths [EMAIL PROTECTED] Heriot-Watt University Edinburgh EH14 4AS, U.K. Phone: +44 131 451 3893 Fax: +44 131 451 3249 Former address: Dept. de Física de la Materia Condensada Facultad de Ciencias, Universidad de Zaragoza 50009 Zaragoza, SPAIN =
Re: NT y LINUX
Hola que tal. Veo que el tema se animo. Lo primero que yo intente fue instalar primero NT y despues LINUX pero al poner LILO en el MBR, NT ya no puede arrancar ya que necesita su propio MBR, es decir el boot loader de NT, me temo que NT usa el boot loader para algo o es una nueva conia de M$ para no facilitar la instalacion de otros sistemas.
Re: NT y LINUX
On mié, abr 07, 1999 at 11:23:30 +0100, Jose Marin wrote: Supongo que lo mejor seria poner LILO en el MBR (i.e., boot=/dev/hda), y Si tenerlo asi de master bootloader. Qué creeis? Pero, en ese caso, sabe alguien como guardar el MBR original (el bootloader de NT), por si interesa dejarlo como estaba en un futuro? Mete un disquete formateado y haz... 'dd if=/dev/hda of=/dev/fd0/MBRwinNT bs=512 count=1' De esta forma si la pifias no tendrás por qué alarmarte, simplemente: arranca con el disco de arranque y reestablece la MBR de WinNT mediante: 'dd if=/dev/fd0/MBRwinNT of=/dev/hda bs=446 count=1' Saludos. -- Javier Viñuales Gutiérrez [EMAIL PROTECTED]
Re: NT y LINUX
El Mon, Mar 29, 1999 at 03:58:17PM +0200, Ramiro Alba dijo: José Enrique Álvarez Martín wrote: [Problemas de Arranque Linux+WinNT] Ya vieron los HOWTO relevantes. -- Ugo Enrico Albarello López de Mesa| POWERED BY | www.debian.org [EMAIL PROTECTED] | DEBIAN GNU/LINUX 2.0 | www.gnu.org - Always Free, Always Cool, Always Linux
Re: NT y LINUX
On Mon, Mar 29, 1999 at 03:58:17PM +0200, Ramiro Alba wrote: Tenemos Windows NT 4.0 instalado en una partición del primer disco y en otra(s) particiones del mismo disco instalamos Debian y onfiguramos Lilo para que arranque de los 2 sistemas. El arranque de Linux ningún problema pero el de NT comienza bien hasta que aparece la pantallita azul y despues de unos 10 segundos falla estrepitosamente. Esto no ocurre si el disrectorio root de Linux esta en una particion de otro discos. Si no es el caso, me he visto obligado a poner el Lilo en disket para el arranque dual. Hasta donde he podido me mirado la documentación de Lilo a fondo, pero no he dado con la causa. ¿Alguien sabe que demonios pasa? Yo lo plantearía de otra manera, que es la que a mi me ha funcionado: instalas lilo en la partición que va contener a linux, siempre por debajo del cilindro 1024 y en disco master. Luego copias el sector de arranque de linux en un fichero, con dd if=/dev/hda2 of=/bootsect.lnx bs=521 count=1 (suponiendo que linux esté en hda2). Lo siguiente es copiar el archivo /bootsect.lnx a c: Si c: es una partición ntfs, tendrás que copiarlo primero en un disco. Entonces editas el archivo boot.ini, que es (hablo de memoria) hidden,read-only,system. Antes de editarlo tendrás que cambiarle los atributos, pero luego acuérdate de dejarlos como estaban. Como decía, editas el fichero y añades la línea c:\bootsect.lnx=Linux Con esto, serás capaz de arrancar linux desde el cargador de NT. Funciona perfectamente (doy fe); el único inconveniente es que cada vez que retocas el lilo, tienes que copiar el sector de arranque de nuevo. Todo esto viene mejor explicado en el howto Linux+NT+loader (creo que se llamaba así) y en el número 5 de Linux Actual. -- Un Saludo Han Solo The Rebel Alliance Conecto, luego existo. Desconecto, luego insisto. Soy usuario de infobirria+ P.D. La firma no es mía, sino de uno que trabajaba, precisamente, en M$. Vivir para ver.
Re: NT y LINUX
José Enrique Álvarez Martín wrote: Hola a todos. Instale Windows NT 4.0 Server en mi pc, en una particion ntfs.Mas tarde instale LINUX en otro disco duro, al instalar LILO, me machaco el MBR de NT pero lo peor es que no puedo arrancar NT desde LILO. Por favor, alguien me puede ayudar. Si, a mi también me paso lo mismo y lo arrglé arrancando desde disket DOS y ejecutando fdisk /MBR (el disket de arranque DOS ha de contener fdisk). Al hilo de la pregunta plantearé otra: Tenemos Windows NT 4.0 instalado en una partición del primer disco y en otra(s) particiones del mismo disco instalamos Debian y onfiguramos Lilo para que arranque de los 2 sistemas. El arranque de Linux ningún problema pero el de NT comienza bien hasta que aparece la pantallita azul y despues de unos 10 segundos falla estrepitosamente. Esto no ocurre si el disrectorio root de Linux esta en una particion de otro discos. Si no es el caso, me he visto obligado a poner el Lilo en disket para el arranque dual. Hasta donde he podido me mirado la documentación de Lilo a fondo, pero no he dado con la causa. ¿Alguien sabe que demonios pasa? Un saludo a todos -- Ramiro Alba Laboratori de Termotecnia i Energetica Departament de Maquines i Motors Termics ETS d'Enginyers Industrials de Terrassa C/Colom 11 Tf: 34 - 93 739 82 43 Fax: 34 - 93 739 81 01 e-mail: [EMAIL PROTECTED]
Re: NT y LINUX
Jos Enrique lvarez Martn wrote: Hola a todos. Instale Windows NT 4.0 Server en mi pc, en una particion ntfs. Mas tarde instale LINUX en otro disco duro, al instalar LILO, me machaco el MBR de NT pero lo peor es que no puedo arrancar NT desde LILO. Por favor, alguien me puede ayudar. Ver documentacion , por ejemplo en http://sunsite.unc.edu/LDP/HOWTO/mini/Linux+NT+Loader Antonio
RE: NT and Linux
Thanks Bob McGowan for your very informative reply. I gather that 1. Software raid is OK if problem is I-O bound, i.e., CPU would normally be idle waiting for I-O. 2. If we have multiple subsystems, we increase the the I-O bandwidth, and now the CPU may not be keep up with the I-O. In general, increasing I-O turns I-O bound problem into CPU bound program. 3. Software raid 5 may be OK for workload with lots of reads, but run into trouble if workload does lots of writes. 4. Software raid 5 is more efficient for large files. Is the above more or less correct. King On Mon, 1 Jun 1998, Bob McGowan wrote: On Thu, 28 May 1998, Leandro Guimaraens Faria Corcete Dutra wrote: snipped The article from www.osnews.com did say that software raid takes up CPU cycles, but it did not say how much. It would seem that if the CPU must check for errors on each byte from disk and performance would take a big hit. Perhaps the kernel checks for errors only if it knows that a disk died, and normally there would not be a hit. Does anyone know about CPU hit of software raid. Why would anyone buy expensive raid hardware if software does the same without too much penalty? King Lee First, the CPU not only checks for errors on reading, it must also calculate the parity on writes. In RAID5, spanning 4 disks, for example, 1/4 of the storage is used to hold parity info. Data is written in stripes of some size, one stripe per disk, in a round robin sequence. One stripe will be parity. In the above 4 disk example, if a stripe were 16K in size, there would be 48K of data and 16K of parity. In RAID5, the parity stipe will rotate between disks, so no single disk is loaded with all the parity (this improves performance over RAID4(I believe) where all parity is on one disk). If a disk write is less than 48K, the system must read 48K from the disks, make the needed changes, recalculate parity and write the resulting 64K back to the disks. If the size is 48K, this read of data can be dispensed with. The system must then only calcualte the parity and then write the 64K. This means CPU cycles are needed for SW RAID. I do not know the impact in terms of actual numbers, but I can say the main issue is scalability. In SW RAID, the more RAID subsystems created, the greater the impact on CPU performance. In HW RAID, there is no additional impact. So even if SW RAID for a single RAID5 subsystem matched HW RAID for the same config, there will certainly come a breakeven point, where additional capacity causes CPU performance degradation in the SW RAID setup. --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: NT and Linux
Hi, King, my comments follow your questions, below. I hope this helps. Bob King Lee asks: Thanks Bob McGowan for your very informative reply. I gather that 1. Software raid is OK if problem is I-O bound, i.e., CPU would normally be idle waiting for I-O. I would agree with this analysis. If the CPU is doing nothing, it might as well be calculating parity for RAID. :-) 2. If we have multiple subsystems, we increase the the I-O bandwidth, and now the CPU may not be keep up with the I-O. In general, increasing I-O turns I-O bound problem into CPU bound program. I would also expect this to be true, though I have no evidence to support the idea. 3. Software raid 5 may be OK for workload with lots of reads, but run into trouble if workload does lots of writes. Not necessarily. Remember, when reading the data, you still have to read a stripe from all the disks and verify the parity, so there is still some overhead. Also, if there are lots of writes, there may be a higher chance of ordering the I/O requests to take advantage of writing a full set of stripes, reducing the frequency of the read/modify/write cycle, which will reduce I/O load. 4. Software raid 5 is more efficient for large files. Generally, the answer to this is: it depends ;-) Are you talking reads and/or writes. What combination? How random? Etc. Also, this question (and the third, to some extent) are getting away from the original question comparing SW and HW based RAID technnology and are getting into the more specific issues of RAID efficiencies, which DO NOT depend on whether the RAID is SW or HW. Generally, in RAID5, writes will always be more expensive than a regular disk. If you have a read/modify/parity calcualtion/write scenario, it is worse, but even the data collection/parity calculation/write sequence takes more time than a pure write. The efficiency of RAID5 is in its read characteristics, for random access. Large numbers of random read requests will distribute across multiple spindles, improving I/O due to redcution of seek delays and an overall reduction of read requests PER SPINDLE. There will also be less wait time for unrelated requests. This implies that the more disks you can put in the array, the better the performance. And this may be where SW RAID could be better than HW RAID, since SW based arrays can span multiple controllers. The controllers also do not need to be the same interface type either. You can mix IDE, SCSI, etc. HW RAID systems generally have some limits on the number of disks you can have, based on the number of internal buses and bus width (ie a two internal narrow SCSI channel system would be limited to a maximum of 14 hard disks). If you are concerned about write performance more than read performance, you might want to consider using a mirror set of some sort (RAID1 and RAID6 [AKA RAID10]). Since there is no parity calculation, write performance is very close to a standard disk's. The disadvanage is that 50% of the capacity is lost. Is the above more or less correct. King On Mon, 1 Jun 1998, Bob McGowan wrote: On Thu, 28 May 1998, Leandro Guimaraens Faria Corcete Dutra wrote: snipped The article from www.osnews.com did say that software raid takes up CPU cycles, but it did not say how much. It would seem that if the CPU must check for errors on each byte from disk and performance would take a big hit. Perhaps the kernel checks for errors only if it knows that a disk died, and normally there would not be a hit. Does anyone know about CPU hit of software raid. Why would anyone buy expensive raid hardware if software does the same without too much penalty? King Lee First, the CPU not only checks for errors on reading, it must also calculate the parity on writes. In RAID5, spanning 4 disks, for example, 1/4 of the storage is used to hold parity info. Data is written in stripes of some size, one stripe per disk, in a round robin sequence. One stripe will be parity. In the above 4 disk example, if a stripe were 16K in size, there would be 48K of data and 16K of parity. In RAID5, the parity stipe will rotate between disks, so no single disk is loaded with all the parity (this improves performance over RAID4(I believe) where all parity is on one disk). If a disk write is less than 48K, the system must read 48K from the disks, make the needed changes, recalculate parity and write the resulting 64K back to the disks. If the size is 48K, this read of data can be dispensed with. The system must then only calcualte the parity and then write the 64K. This means CPU cycles are needed for SW RAID. I do not know the impact in terms of actual numbers, but I can say the main issue is scalability. In SW RAID, the more RAID subsystems created, the greater the impact on CPU performance.
Re: NT and Linux
On Fri, 29 May 1998, Michele Comitini wrote: One great advantage is that you can combine any kind of partitions form different devices (even a combination of partitions from a mix of IDE or SCISI hard-disks!) and have different personalities (i.e. RAID-5 for filesystem partitions, RAID-0 for swap partitions) on partitions of the same hard-disk. Note that you don't need RAID0 to do striping on swap partitions. You can assign each swap partition a priority. If all have the same priority, the kernel (versions 1.3.6 and higher) will automatically use something like striping on them. For more info, see 'man 8 swapon' and 'man 2 swapon' for more info. Remco -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: NT and Linux
-Original Message- From: King Lee [mailto:[EMAIL PROTECTED] Sent: Thursday, May 28, 1998 11:29 PM To: Leandro Guimaraens Faria Corcete Dutra Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: Re: NT and Linux On Thu, 28 May 1998, Leandro Guimaraens Faria Corcete Dutra wrote: King Lee wrote: 1. Has anyone here had any experience or knowledge about software raid. How good is it? 2. Does Linux support hardware raid 5 snipped The article from www.osnews.com did say that software raid takes up CPU cycles, but it did not say how much. It would seem that if the CPU must check for errors on each byte from disk and performance would take a big hit. Perhaps the kernel checks for errors only if it knows that a disk died, and normally there would not be a hit. Does anyone know about CPU hit of software raid. Why would anyone buy expensive raid hardware if software does the same without too much penalty? King Lee First, the CPU not only checks for errors on reading, it must also calculate the parity on writes. In RAID5, spanning 4 disks, for example, 1/4 of the storage is used to hold parity info. Data is written in stripes of some size, one stripe per disk, in a round robin sequence. One stripe will be parity. In the above 4 disk example, if a stripe were 16K in size, there would be 48K of data and 16K of parity. In RAID5, the parity stipe will rotate between disks, so no single disk is loaded with all the parity (this improves performance over RAID4(I believe) where all parity is on one disk). If a disk write is less than 48K, the system must read 48K from the disks, make the needed changes, recalculate parity and write the resulting 64K back to the disks. If the size is 48K, this read of data can be dispensed with. The system must then only calcualte the parity and then write the 64K. This means CPU cycles are needed for SW RAID. I do not know the impact in terms of actual numbers, but I can say the main issue is scalability. In SW RAID, the more RAID subsystems created, the greater the impact on CPU performance. In HW RAID, there is no additional impact. So even if SW RAID for a single RAID5 subsystem matched HW RAID for the same config, there will certainly come a breakeven point, where additional capacity causes CPU performance degradation in the SW RAID setup. --- Bob McGowan i'm: bob dot mcgowan at artecon dot com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: NT and Linux
On Thu, 28 May 1998, Leandro Guimaraens Faria Corcete Dutra wrote: King Lee wrote: 1. Has anyone here had any experience or knowledge about software raid. How good is it? 2. Does Linux support hardware raid 5 Just (re)found it! http://www.osnews.com./features/04.98/raid.html Very good reading indeed! Enjoy and tell us what has come of it! Thanks for info. Also http://www.linas.org/linux/raid.html had some very good info. I was surprised to learn that the 2.2 kernel supports software raid and that the software raid was as fast as hardware raid 5. Raid 5 does error correction and even if one of the disks die data can be recovered and the system continue. The article from www.osnews.com did say that software raid takes up CPU cycles, but it did not say how much. It would seem that if the CPU must check for errors on each byte from disk and performance would take a big hit. Perhaps the kernel checks for errors only if it knows that a disk died, and normally there would not be a hit. Does anyone know about CPU hit of software raid. Why would anyone buy expensive raid hardware if software does the same without too much penalty? King Lee -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: NT and Linux
Hello! I was surprised to learn that the 2.2 kernel supports software raid and that the software raid was as fast as hardware raid 5. Raid 5 does error correction and even if one of the disks die data can be recovered and the system continue. The article from www.osnews.com did say that software raid takes up CPU cycles, but it did not say how much. It would seem that if the CPU must check for errors on each byte from disk and performance would take a big hit. Perhaps the kernel checks for errors only if it knows that a disk died, and normally there would not be a hit. Does anyone know about CPU hit of software raid. Why would anyone buy expensive raid hardware if software does the same without too much penalty? King Lee Well as a matter of fact I realized a Debian system with the software RAID-5 almost one year ago and it had good performance. Anyway I have never done any serious performance testing on it. The big problem is having the whole filesystem under RAID-5 even the root filesystem, this was solved using the initrd ramdisk to activate the RAID-5 personality on the partitions selected. This was probably the biggest problem with the linux software RAID. One great advantage is that you can combine any kind of partitions form different devices (even a combination of partitions from a mix of IDE or SCISI hard-disks!) and have different personalities (i.e. RAID-5 for filesystem partitions, RAID-0 for swap partitions) on partitions of the same hard-disk. I do not think you can do the same with a RAID-5 capable controller. After all it is probably cheaper and more effective to buy a dual (or quad) CPU motherboard instead of buying an expensive controller, but you have to do much more work on your side. Best Regards, Michele Comitini -- E-Mail: Michele Comitini [EMAIL PROTECTED] Date: 29-May-98 Time: 10:41:23 This message was sent by XFMail -- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: NT and Linux
-Original Message- From: King Lee [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 27, 1998 8:29 PM To: debian-user@lists.debian.org Subject: NT and Linux Hello, I got into a discussion with a system administrator of a website. The system administrator wishes to use NT because it supports software raid 5 (raid without a special controller). I thought if it works, there would be a terrible performance degradation. The system administrator said only if a disk goes down would there be a performance hit. Does anyone here know anything about The questions I have are 1. Has anyone here had any experience or knowledge about software raid. How good is it? 2. Does Linux support hardware raid 5 I think this guy is looking for an excuse not to use Linux. King Lee 1. I'm currently got a bundle of Alpha-Servers running NT 4.0. All but one of them uses hardware raid 5 (Controlled by an HSZ40 in a DEC storage works cab.) It is DEFINATELY faster than the one which has NT controlled RAID 5. Hardware controlled software controlled sets are both same sizes on each server. This difference is phenomenal. Aside from mere disk access - you should see how much CPU time is used keeping the RAID set working on the software controlled one.. I'll never setup software RAID again. 2. If Linux can see SCSI disks (which it can) it's all just a matter of plugging your SCSI cable into a hardware RAID box (such as the Digital Storage Works cabinets). The hardware controller takes care of everything. Linux will just see it as one huge disk. You will however need to spend an extra half-an-hour setting up the controller (It'll be done through a dumb terminal - and it's dead easy). Robbie e-mail: [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: NT and Linux
King Lee wrote: 1. Has anyone here had any experience or knowledge about software raid. How good is it? 2. Does Linux support hardware raid 5 Just (re)found it! http://www.osnews.com./features/04.98/raid.html Very good reading indeed! Enjoy and tell us what has come of it! -- Leandro Guimaraens Faria Corcete Dutra http://www.lge.com.br./ [EMAIL PROTECTED] http://www.terravista.pt./Enseada/1989/ BRASIL -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: NT and Linux
King Lee wrote: 1. Has anyone here had any experience or knowledge about software raid. How good is it? Know nothing about NT. If you look for information on Linux RAID (it's in the Internet, I've read it, can't remember where), it's said that Linux s/w RAID was in fact proved to be faster than the h/w products. 2. Does Linux support hardware raid 5 Yes! I think this guy is looking for an excuse not to use Linux. Do not be hard on him. I was avoiding Unix too, until I experienced NT's failures... -- Leandro Guimaraens Faria Corcete Dutra http://www.lge.com.br./ [EMAIL PROTECTED] http://www.terravista.pt./Enseada/1989/ BRASIL -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
RE: NT and Linux
-Original Message- From: King Lee [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 27, 1998 12:29 PM To: debian-user@lists.debian.org Cc: recipient list not shown; @[EMAIL PROTECTED] Subject: NT and Linux Hello, I got into a discussion with a system administrator of a website. The system administrator wishes to use NT because it supports software raid 5 (raid without a special controller). I thought if it works, there would be a terrible performance degradation. The system administrator said only if a disk goes down would there be a performance hit. There will be some performance loss, since the system CPU will need to handle the RAID algorithm. Software RAID also means buying the SW to support it or having it come with the system (as it does with NT). It is still necessary to purchase the disk farm and perhaps more HBA's to distribute the load and improve redundancy. There are no restrictions that I know of about the interface or disk types used. One issue is how much I/O is written at a time and the size of the stripes of data written to each disk. As an example a 5 disk array using a stripe size of 16K will have a big performance hit (whether implemented in SW or HW) if a write of less than 4*16K (64K) is made. The RAID system must then read the unchanged data from the disks, make the needed changes, calculate the new parity and write the whole thing back. This takes CPU cycles and will affect system performance at some load levels. And even if the writes are full stripes, it still needs to calculate the parity and write the stripes to disk. Another issue is the type of I/O being done (random vs. sequential) but this impacts I/O performance in either SW or HW RAID. The more random the I/O, the beter the chances are that several writes (or reads) will land on different disks in the array, reducing seek time issues. I/O performance will also improve as more threads are run. Does anyone here know anything about The questions I have are 1. Has anyone here had any experience or knowledge about software raid. How good is it? I have used it, but not recently and not in a production environment. It does/did work. (I'm a test engineer so I beat the hell out of it. I had no failures or problems.) 2. Does Linux support hardware raid 5 Basically, any system can support hardware RAID at any level, since the RAID functions are handled by the RAID controller. But then there needs to be some way to configure the RAID subsystem. This can be done by either a serial interface to the RAID subsystem controller, using a terminal emulator, or by special software using a SCSI pass through to send information to the controller over the SCSI bus. This assumes a SCSI subsystem of course. The subsystem manufacturers are building high performance systems, so the dollar outlay can be large (4 or 9 GB 7200 RPM Ultra SCSI disks in a cabinet running Ultra SCSI to the host, supporting a large number of drives (7 or more)). The serial method is fast and easy but does not scale well to large numbers of systems, where the SCSI base scales nicely but is more difficult to implement well. I think this guy is looking for an excuse not to use Linux. King Lee -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: NT vs. Linux: is zero-administration a reality? (was: Question.)
Jens B. Jorgensen [EMAIL PROTECTED] writes: [...] NT make simple things simpler. In the process, by cramming everything into a neat little GUI it makes complex things difficult or impossible. Amen, brother! Truer words were never spoken. Well, maybe now and then, but not often. :) -- Edgar Whipple Have clue, will travel. [EMAIL PROTECTED]Budgies?! We doan need no stinkin *budgies*!! Microsoft is not where I want to go today. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .