Re: Help with hardware choices
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Wed, 30 Mar 2005, Brandon Mercer wrote: Also, I don't know how you'd ever build a box for 700 bucks running scsi. Actually, in the comparison Dave was making (Sun Blade 1500 to an Opteron based sever), SCSI probably isn't applicable as a Sun Blade 1500 ships with IDE drives by default. Obviously if you are thinking later Sun Netra or Sun Fire models, then that's a whole different story. Cheers, - -- Jason Wever Gentoo/Sparc Co-Team Lead -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFCS2D8dKvgdVioq28RAoHuAJ9QTb0QPjABYDNXu548yGN39KO8qQCbBTKV QnEl0IleNZW6qETaMl1l88w= =HDfy -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: __alloc_pages: 2-order allocation failed
On Wed, 30 Mar 2005 19:01:17 -0500 Mike Edwards <[EMAIL PROTECTED]> wrote: > The reason I asked was because the IBM article I linked to earlier > seemed to indicate the problem cropped up on 64 bit systems. That problem in that article is S390 platform specific. They use these higher-order allocations for the S390 page tables. Sparc64 only uses up to order 1 allocations for it's page tables so wouldn't trigger the bug you saw. > In any case, I can't recall coming across this issue before, and most > of the systems I run have far less memory than this beast. :) > > Is there a way to fix this? Try to find out what is asking for such a large allocation. You can do that by applying a patch similar to this one so that the debug message prints out more information: = mm/page_alloc.c 1.72 vs edited = --- 1.72/mm/page_alloc.c2004-08-08 01:58:48 -07:00 +++ edited/mm/page_alloc.c 2005-03-30 16:21:23 -08:00 @@ -476,6 +476,8 @@ out: printk(KERN_NOTICE "__alloc_pages: %u-order allocation failed (gfp=0x%x/%i)\n", order, gfp_mask, !!(current->flags & PF_MEMALLOC)); + printk(KERN_NOTICE "__alloc_pages: task(%s) pid(%d) caller(%p)\n", + current->comm, current->pid, __builtin_return_address(0)); if (unlikely(vm_gfp_debug)) dump_stack(); return NULL; -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Problem with /tmp d-i BUG?
After my succesfull install of sarge on a Balde 100 I doscoverd that the permissions for /tmp are wrong. It should be drwxrwxrwx but it is drwxr-xr-x also the partition recipe of d-i created a seperated tmp partition /dev/hda7 there was no ext3 filesystem created on it and /dev/shm tmpfs was mounte on the /tmp moint point anyway. /etc/fstab however has the right line for it /dev/hda7 /tmpext3defaults0 2 looks like that tmpfs behavier is triggered by the existence of the /etc/default/tmpfs file putet there by the initscripts package. And on other thing mkfs -t fstype gives only a not found. mkfs.ext3 and so on are were they should be. Any Idea to what packets I should file bug reports or why that behavier didn't show up on my i386/AMD64 boxes? greets Uwe -- Jetzt will man das Internet nicht einfach ein paar Leuten wie der IETF überlassen, die wissen, was sie tun. Es ist zu wichtig geworden. - Scott Bradner http://www.highspeed-firewall.de/adamantix/ http://www.x-tec.de
Re: __alloc_pages: 2-order allocation failed
The reason I asked was because the IBM article I linked to earlier seemed to indicate the problem cropped up on 64 bit systems. In any case, I can't recall coming across this issue before, and most of the systems I run have far less memory than this beast. :) Is there a way to fix this? On Wed, Mar 30, 2005 at 03:17:24PM -0800, David S. Miller said: > On Wed, 30 Mar 2005 17:58:58 -0500 > Mike Edwards <[EMAIL PROTECTED]> wrote: > > > This doesn't sound useful for servers. :) Is this an issue with Linux > > in general, or only with certain ports? > > When I said "The kernel" I did mean Linux in general not for a > specific port such as Sparc. -- Mike Edwards <[EMAIL PROTECTED]> System Administrator Psychology Department, Rutgers University, Newark campus -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Help with hardware choices
On Wed, 30 Mar 2005 18:20:25 -0500 Brandon Mercer <[EMAIL PROTECTED]> wrote: > I do not know what you're talking about. Sparcs use considerably less > power, run cooler, last longer, are more stable and are comparable with > speed. You've got to be kidding me. Maybe for ancient pre-UltraSPARC boxes, but my dual-750Mhz SunBlade1000 alone I can hear humming in the very next room and I don't need any heat in that room because of that machine. Similarly for my Ultra60. > Don't confuse clock cycles with speed either. I'm not at all. I've been maintaining and working on the sparc Linux ports for at least 10 years. I rewrote the GCC compiler backend for Sparc from scratch. I wrote all of the instruction scheduling descriptions in the SPARC backend of GCC. I'm the only person who has bothered trying to write assembler optimized code for UltraSPARC in libmpeg and MESA. It might be possible that I kind of know what I'm talking about when I say that SPARCs simply aren't a good bang for the buck these days. I know the hardware inside and out, so given that do you think I'm qualified to make those statements? > I've got an old sparcstation 20 with a 60Mhz proc, that performs just as > well as my Pentium200 for network routing. Yes, hardware from back in the mid-80's, the glory days of Sun hardware when it actually was comparable to the x86 offerings of the same time frame. Some things have changed in the last 15 years :-) I have a strong vested interest in SPARC doing well, yet I can sit here and say SPARC really has lost the cpu wars. I wish other Sun/SPARC lovers could be similarly realistic. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: __alloc_pages: 2-order allocation failed
On Wed, 30 Mar 2005 17:58:58 -0500 Mike Edwards <[EMAIL PROTECTED]> wrote: > This doesn't sound useful for servers. :) Is this an issue with Linux > in general, or only with certain ports? When I said "The kernel" I did mean Linux in general not for a specific port such as Sparc. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Help with hardware choices
David S. Miller wrote: >On Wed, 30 Mar 2005 15:51:29 -0600 >Kent West <[EMAIL PROTECTED]> wrote: > > > >>Dale Scheetz wrote: (snippage throughout) >> >> >> >>>I have been tasked with picking the hardware, and my feeling is that >>>SPARC offers the best bang for the buck. >>> >>> >>Let me preface this by saying I have no idea what I'm talking about, but >>if it were me, I'd at least take a look at the Mac-mini (this site is >>doing web hosting on Minis - >>http://www.xrackhosting.com/machine.php?pid=dedicated_hosting) and/or >>the Xserve if you need that much horsepower. >> >> > >One thing is for sure, SPARC is definitely not the best bang for >the buck. > >Case in point, I just dropped $2600.00USD on a new SunBlade1500 with 1.5GHZ >cpus and 1GB of ram. I know that for around $700.00USD in parts you can put >together your own Opteron system which is twice as powerful. > >The SunBlade sucks twice as much power and is twice as loud as the Opteron >as well. The SPARC machines only makes sense for people deeply interested >in Sparc already, such as myself. :-) > > > I do not know what you're talking about. Sparcs use considerably less power, run cooler, last longer, are more stable and are comparable with speed. Don't confuse clock cycles with speed either. Also, I don't know how you'd ever build a box for 700 bucks running scsi. All I'm saying is that you're not comparing apples to apples. I've got an old sparcstation 20 with a 60Mhz proc, that performs just as well as my Pentium200 for network routing. And I know for a fact it uses less power. Brandon -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: __alloc_pages: 2-order allocation failed
This doesn't sound useful for servers. :) Is this an issue with Linux in general, or only with certain ports? Is there a patch, or something else I can do to keep this from happening again? As far as uptime goes... 17:58:45 up 78 days, 5:08, 18 users, load average: 0.28, 0.27, 0.23 On Tue, Mar 29, 2005 at 12:07:04PM -0800, David S. Miller said: > On Tue, 29 Mar 2005 13:43:10 -0500 > Mike Edwards <[EMAIL PROTECTED]> wrote: > > > The odd thing is, psychology shouldn't be running out of memory - this > > beast has 4 GB. > > It's not running out of memory so much as memory is too fragmented > to produce a 2-order allocation successfully. > > The kernel uses the buddy-system algorithm to allocate free pages, > this means when a 2-order allocation is requested the kernel has to > find 4 consequetive and aligned pages in the free page pool. Sometimes > after lots of uptime, this is difficult if not impossible to achieve. -- Mike Edwards <[EMAIL PROTECTED]> System Administrator Psychology Department, Rutgers University, Newark campus -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Help with hardware choices
On Wed, 30 Mar 2005 15:51:29 -0600 Kent West <[EMAIL PROTECTED]> wrote: > Dale Scheetz wrote: (snippage throughout) > > >I have been tasked with picking the hardware, and my feeling is that > >SPARC offers the best bang for the buck. > > Let me preface this by saying I have no idea what I'm talking about, but > if it were me, I'd at least take a look at the Mac-mini (this site is > doing web hosting on Minis - > http://www.xrackhosting.com/machine.php?pid=dedicated_hosting) and/or > the Xserve if you need that much horsepower. One thing is for sure, SPARC is definitely not the best bang for the buck. Case in point, I just dropped $2600.00USD on a new SunBlade1500 with 1.5GHZ cpus and 1GB of ram. I know that for around $700.00USD in parts you can put together your own Opteron system which is twice as powerful. The SunBlade sucks twice as much power and is twice as loud as the Opteron as well. The SPARC machines only makes sense for people deeply interested in Sparc already, such as myself. :-) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Help with hardware choices
In article <[EMAIL PROTECTED]> you write: >I have been tasked with picking the hardware, and my feeling is that >SPARC offers the best bang for the buck. I have been pointed at two >different lines of machines by a Systems Engineer for Sun here in >Tallahassee, but I was hoping that you guys would have a better idea >of what I can expect from these machines. > >Obviously it must run the Debian SPARC port, and while the Sun webpage >lists Linux as a supported OS, they reference things like Red Hat >Enterprise, which doesn't make it clear whether Debian is supported. >(I mean to have a talk with them about that as well ;-) Sun makes good reliable, but not cheap, hardware. However, I would not recomend getting a new Sparc system to run Debian. Sparc is one of the architectures that may not make it into etch (the release after the upcoming sarge) and is generally not as well supported as x86 machines. Debian sparc developers tend not to have the latest hardware, so there may have been no testing on your configuration. For the stated application, I'd look at server class Opteron (AMD64) or Xeon systems. You need to trade off absolute specs for reliability, and don't forget to look at having a backup system as way of gaining that reliablility. -- Blars Blarson [EMAIL PROTECTED] http://www.blars.org/blars.html With Microsoft, failure is not an option. It is a standard feature. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Help with hardware choices
Dale Scheetz wrote: (snippage throughout) I have been working for the Florida Department of State, Division of Cultural Affairs. I am currently building a new database for the Division. It is web based Apache/php4/postgreSQL software that will allow Individuals and Organizations to view and update their personal information as well as view the status of their various grants and provide staff with view/edit screens to access the complete data set. I have been tasked with picking the hardware, and my feeling is that SPARC offers the best bang for the buck. Let me preface this by saying I have no idea what I'm talking about, but if it were me, I'd at least take a look at the Mac-mini (this site is doing web hosting on Minis - http://www.xrackhosting.com/machine.php?pid=dedicated_hosting) and/or the Xserve if you need that much horsepower. -- Kent West Technology Support /A/bilene /C/hristian /U/niversity -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Proposal to document sparc installation issues
On Wednesday 30 March 2005 14:07, Holger Levsen wrote: > On Tuesday 29 March 2005 21:19, Frans Pop wrote: > > The first issue is with USB keyboards by Sun as used on > > for example SunBlade systems. These are incorrectly "recognized" by > > the installer as regular Sun keyboards. A workaround is documented in > > the Installation Guide (see link above, chapter "Using the Debian > > Installer"). > > Does this happen with 2.6 or 2.4 or both ? Only with 2.4; for 2.6 kernels all keyboards are configured with "PC" keymaps. > > The second issue is kernel related. Kernels in the 2.6 > > series use a different input layer that makes all keyboards look like > > "normal" PC keyboards. This means that if you boot the installer with > > a 2.4 kernel and configure it for a Sun or USB keyboard and later (in > > expert mode) select a 2.6 kernel for the new system, you will very > > likely end up with a non-working keyboard after reboot. > > Can this be workaround by "dpkg-reconfigure console-data" before the > reboot ? Unfortunately not, as setting up the keyboard is one of the very last things to be done before the reboot. The best option here is to just select the same kernel version as the installer was booted with (as happens automatically during a default install). I will try to make this a bit clearer in my texts, but I don't want to make the texts too long. pgp0X82LFsq70.pgp Description: PGP signature
Help with hardware choices
For the past several years I have been working for the Florida Department of State, Division of Cultural Affairs. We make grants to Florida Arts organizations and some small stipends to Individual Artists. I am currently building a new database for the Division. It is web based Apache/php4/postgreSQL software that will allow Individuals and Organizations to view and update their personal information as well as view the status of their various grants and provide staff with view/edit screens to access the complete data set. The machine I am doing the development on is a crumby slow Gateway machine, and we new from the start that this would not be our server. I have been tasked with picking the hardware, and my feeling is that SPARC offers the best bang for the buck. I have been pointed at two different lines of machines by a Systems Engineer for Sun here in Tallahassee, but I was hoping that you guys would have a better idea of what I can expect from these machines. Obviously it must run the Debian SPARC port, and while the Sun webpage lists Linux as a supported OS, they reference things like Red Hat Enterprise, which doesn't make it clear whether Debian is supported. (I mean to have a talk with them about that as well ;-) So, after that long-winded introduction, we get to my question: The machines I've been looking at are the Xeon V65X and the Ultra V210, V240, and possibly th;e v490. My question is: Which of these machines would best serve my needs as a web server/SQL server, and are there any other machines that would suite my needs better. Obviously the more horsepower the better, but I'm also on a budget of 7 to 10 thousand dollars, so the quad Xeon (at $20,000) is probably out of the question. My feeling is that the SPARC is the machine I want, but my only basis for this decission is that I own one (BTW what ever happend to Kachina Tech?). All suggestions greatfully accepted. Please CC me as I am not subscribed to this list. Waiting is, Dwarf -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: can't partition disk?
* Uwe A. P. Wuerdinger <[EMAIL PROTECTED]> [2005-03-30 03:27]: > >>Is this on O2? It worked for me last night. > >Yep. I just went through it again. I wonder if the install gets > >confused if it thinks there's already a previous installation? > > I've seen exactly that behavier with RC3 netboot.img on a > SUN Blade 100. On Sun? How weird. > Do you have any logs to open up a bug? When this happens, can you go to the main menu and open a shell (the last but one entry), and then go to /var/log/ and see if there's any information (most probably in messages). -- Martin Michlmayr http://www.cyrius.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
AVVISO DI VIRUS / VIRUS WARNING : Worm.SomeFool.P
Il messaggio sotto riportato, spedito dal vostro indirizzo di posta, contiene un VIRUS e pertanto non è stato consegnato. Probabilmente il computer dal quale è stato spedito è infetto. CONTROLLATELO QUANTO PRIMA CON UN PROGRAMMA ANTIVIRUS! A message containing a virus was sent from your e-mail address. It is very likely this machine (or any other you use for e-mail) is infected! CHECK IT AS SOON AS POSSIBLE WITH AN ANTIVIRUS PROGRAM! Received: (qmail 12924 invoked from network); 30 Mar 2005 09:06:09 - Received: from unknown (HELO igorcognolato.com) (82.104.251.197) by 0 with SMTP; 30 Mar 2005 09:06:09 - From: debian-sparc@lists.debian.org To: [EMAIL PROTECTED] Subject: Mail Delivery (failure [EMAIL PROTECTED]) Date: Wed, 30 Mar 2005 10:52:09 +0200 MIME-Version: 1.0 Content-Type: multipart/related; type="multipart/alternative"; boundary="=_NextPart_000_001B_01C0CA80.6B015D10" X-Priority: 3 X-MSMail-Priority: Normal -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]