- Discuss - lists.centos.org

Autorollback Patches for CentOS 4 RPM
by Sam Hart 07 Jun '05

07 Jun '05

Hi all, For anyone interested, I have just ported James Oden's Autorollback/Autorollback-goal RPM patches to the version of RPM present inside of CentOS 4. The patched RPMs, SRPM and raw patches can be found online here: http://hackers.progeny.com/~sam/RPM-Autorollback/ Specifically, you can find those for CentOS 4 here: http://hackers.progeny.com/~sam/RPM-Autorollback/#wherecentos4 One of the goals of the AR/ARBG patches is to prevent failed RPM transactions from leaving a system in an unusable state. They were designed with Carrier Grade Linux in mind, but I'd wager there could be applications elsewhere. These are the patches taken from RPM's CVS and ported into the CentOS 4 version (4.3.3). Any questions, comments, feel free to send them to me off-list. -- ''''''''''''''''''''''''' .O. Sam Hart, sam(a)progeny.com ..O Progeny Linux Systems, Inc OOO <http://www.progeny.com/>

2 1

Temperature monitoring tools
by William Warren 07 Jun '05

07 Jun '05

Are there any easy to use snmp(mainly temperature monitoring) tools around for Centos -4? This is on a Dell sc420 server. -- Computer House Calls, Networks, Security, Web Design: http://www.emmanuelcomputerconsulting.com What businesses are in Brunswick, Maryland? Check Brunswick First! http://www.checkbrunswickfirst.com My "Foundation" verse: Isa 54:17 No weapon that is formed against thee shall prosper; and every tongue that shall rise against thee in judgment thou shalt condemn. This is the heritage of the servants of the LORD, and their righteousness is of me, saith the LORD. -- carpe ductum -- "Grab the tape" CDTT (Certified Duct Tape Technician) Linux user #322099 Machines: 206822 256638 276825 http://counter.li.org/

5 5

dragging windows leaves traces
by Robin Mordasiewicz 07 Jun '05

07 Jun '05

I have an excellent nvidia graphics card, I get excellent fps, all agp features are installed correctly etc. When I drag windows around wit hthe opaque settings on it leaves traces for a couple milliseconds as it redraws the window, but it just looks so crummy to see. I am having a hard time convincing people that using linux is a good idea. People make judgements on this type of thing and they think that linux is slow compared to Windows. When I had windows installed on this same machine dragging windows around did not produce any traces. Using linux on the desktop looks like a throwback to windows 95 on a 486. Does anyone have any advice on how to make X redraw windows better. --

4 6

RE: [CentOS] Temperature monitoring tools
by Rudi Starcevic 07 Jun '05

07 Jun '05

Hi, >> Are there any easy to use snmp(mainly temperature monitoring) >> tools around for Centos -4? This is on a Dell sc420 server. I use 'lm-sensors' on the assorted servers I admin. Things like CPU temp, FAN speed, Volts etc. Works on dual CPUs nicely too. You may need to compile a kernel with support for the on-board sensor chip. Use 'sensors-detect' to find out. I think you can then use 'sensord' to rig up mrtg or something similar. HTH. Regards, Rudi. -- -- Internet Media Productions

1 0

Gnu Gadu 2
by Maciej Żenczykowski 06 Jun '05

06 Jun '05

Hi, if anyone needs Gnu Gadu 2 [Instant Messaging Program] for Centos 4, I've compiled it (some packages from dag, possibly dries) and it's up at: http://tcs.ii.uj.edu.pl/~maze/centos/gnu-gadu-2/ Cheers, MaZe.

1 0

Re: [CentOS] Re: dragging windows leaves traces -- running older GeForce drivers on newer cards
by Bryan J. Smith <b.j.smith＠ieee.org> 06 Jun '05

06 Jun '05

From: Robin Mordasiewicz <robin(a)bullseye.tv> > [root@zephyr config]# cat /proc/driver/nvidia/agp/status > Status: Enabled > Driver: NVIDIA > AGP Rate: 4x > Fast Writes: Enabled > SBA: Enabled > I get more than 9000FPS running glxgears so the card is doing well. Hmmm, interesting to see both Fast Writes and Side Band Addressing working. Traditionally nVidia has been very conservative. Maybe we'll see a larger performance boost yet! Is this the new 1.0-7xxx series driver? If so, which revision? > I mean by showing the window contents while dragging. As another poster said, if it's the brand new 1.0-7xxx series driver, there were reported performance issues with moving windows. Have you updated to 1.0-7664 (past buggy 1.0-7174)? [ 7664? Wow, that's a lot of builds in a short period of time! ] -- Bryan J. Smith mailto:b.j.smith@ieee.org

2 1

yum setup help and extra repos including centosplus
by Robert Hanson 06 Jun '05

06 Jun '05

greetings recently i asked a question and i was directed to the possibility that a certain fully enabled compiled kernel for CentOS 4 might help me. so, i am looking at http://ftp.osuosl.org/pub/centos/4/centosplus/Readme.txt and wondering how to cover all bases.... can i cover all bases if i put an *blanket* package(s) exclude in /etc/yum.conf ? or do i have to put package(s) exclude in every config file in the /etc/yum.repos.d/ directory ? specifically, the CentOS-Base.repo file. do i put it at the top and then it will be in effect for all sub-repos in that file?? at the top of the CentOS-Base.repo only? :) i realize that i can do it as a command line switch yet then my brains will fall out trying to remember everything. regards and tia for any clarification. - rh

3 4

rpmbuild and stripping libraries
by Maciej Żenczykowski 06 Jun '05

06 Jun '05

Hi, I've just completed a rebuild of php and the rpms came out a _lot_ bigger than the originals... after comparisons it seems stuff isn't being stripped, a comparison of sections on one library reveals the following: > 21 .gnu_debuglink < 21 .comment < 22 .debug_aranges < 23 .debug_pubnames < 24 .debug_info < 25 .debug_abbrev < 26 .debug_line < 27 .debug_frame < 28 .debug_str < 29 .debug_ranges < tons of symbols I'm sure someone knows how to get a normal small build? Pointers? Cheers, MaZe. PS. I used rpmbuild --rebuild php.src.rpm

2 2

PHP mcrypt
by Maciej Żenczykowski 06 Jun '05

06 Jun '05

Hi, Could someone tell me how I'd go about compiling php with mcrypt support from src.rpm? ie. configure -with-mcrypt[=DIR] Is it just a matter of calling rpmbuild --rebuild src.rpm --with=mcrypt ? Cheers, MaZe.

4 7

Grisbi 0.5.7
by Aph 06 Jun '05

06 Jun '05

Grisbi is a personnal accounting application released under the GPL licence. Grisbi RPMS version 0.5.7 for CentOS are now available on our download page http://www.grisbi.org/download.en.html#centos. Regards. -- Ce que je comprends pas, c'est pourquoi leurs votes compte plus ? Et surtout, ca leur rapporte QUOI ???? Putain faut vraiment être CON, grave. -+- A in GNU - La Cabale, c'est plus FORT que toi ! -+-

2 1

Re: dragging windows leaves traces -- running older GeForce drivers on newer cards
by Bryan J. Smith <b.j.smith＠ieee.org> 06 Jun '05

06 Jun '05

From: Robin Mordasiewicz <robin(a)bullseye.tv> > I have an excellent nvidia graphics card, I get excellent fps, all agp > features are installed correctly etc. In Windows or Linux? And by "all agp features are installed correctly," what do you mean and how are you getting this? > When I drag windows around wit hthe opaque settings on it leaves traces > for a couple milliseconds as it redraws the window, but it just looks so > crummy to see. Hmmm, sounds like a window manager/environment issue, unless you're having a video card issue (see below). What is your desktop environment? > I am having a hard time convincing people that using linux is a good idea. Umm, what Windows version offers translucent settings comparable to Linux desktops? None that I know of. And by "opaque," do you mean translucent/ "see through" is turned on, but set to 0%? If so, that's still transcluent. No shipping Windows or Linux distribution I know of uses video card framebuffer to off-load such complex rendering. It's done 100% in software, hence why Microsoft doesn't bother to offer it, and Linux desktops are slow at it. The only major OS I know of that ships using OpenGL to manage window framebuffer directly in video framebuffer is MacOS X via "QuartzExtreme." Microsoft was adding this capability to NT6.0 "Longhorn" in its new, Avalon desktop, but the development has been pushed back and out of NT6.0 "Longhorn" client. To put a "good face" on this, Microsoft is still releasing NT6.0 "Longhorn" client with this next generation Windows Graphics Foundation (WGF), but only version 1.0 which uses existing DirectX 9 code. In a nutshell, DirectX 9 still doesn't have the full geometry setup of OpenGL, which makes it far, far less capable of what QuartzExtreme is capable of with age-old OpenGL 1.2+. WGF 2.0, which incorporates many new features to address these deficiencies (in a nutshell, what was DirectX 10) will ship at a latter date and supposedly address WGF 1.0's shortcomings (I'll believe it when I see it -- there's still a lot of legacy GDI issues to overcome). For more on this, see my PC_Support post here: http://lists.leap-cf.org/pipermail/pc_support/2005-May/000333.html Now in the Linux world, The FreeDesktop.org "Cario" project (not to be confused with the original NT4.0 release codename) adds many things, including direct window-to-video framebuffer via OpenGL over X11 (GLX). It iss being adopted by a number of desktop suites, including Enlightment R17 as well as GNOME 3. A few window managers already implement it, or their own solution. Sun's LookingGlass desktop actually uses a non-X11 system (and quite a number of other, innovative APIs -- including 3D _input_ as _standard_), but offers X11/GLX for compatibility (and, thus, should support "Cario" interfaces as well). But until these features ship as standard in either Windows or Linux distros (think Fedora Core 5 / RHEL 5 / CentOS 5 time-frame ;-), don't expect good performance with opaque desktops where the software is doing really slow and ugly memory maps to/from the video card, over the system interconnect, etc... I.e., don't blame Linux for attempting to implement a feature in software that Microsoft won't care attempt because it has the same issues. ;-> > People make judgements on this type of thing and they think that > linux is slow compared to Windows. > When I had windows installed on this same machine dragging > windows around did not produce any traces. When you say opaque, do you mean translucent but 0% in Linux? Sometimes people enable this feature not realizing that you're adding a massive amount of overhead to the software framebuffer of the window manager. > Using linux on the desktop looks like a throwback to windows 95 on a 486. > Does anyone have any advice on how to make X redraw windows better. Turn off any translucent features, period. You don't want them enabled. From: Sukru TIKVES <sukru(a)cs.hacettepe.edu.tr> > Are you using the x.org driver (nv) or the nvidia one? I'm going to guess the former, or possibly an earlier version of the latter. Unless he has some translucent settings enabled. It would be nice to know what is exactly enabled. From: Robin Mordasiewicz <robin(a)bullseye.tv> > I am using the nvidia driver, with XFree86 centos3.4 > Moving windows is mostly smooth, not all applications leave traces as X > redraws, but I have a dual 3gig machine with a top nvidia card. > Also when resizing windows it seems to redraw much slower than windows. What exact video card on what exact nVidia driver release? Although nVidia has maintained a level of upward compatibility since the original NV0x series (Riva TNT/TNT2) up to the latest NV4x (GeForce 6000 series), if you load an older driver that was never designed for the latest series, you will get similiar symtopms. Example 1: NV2x/NV30-era XFree86 (FC1) with NV40 card Last summer I had upgraded another system, a dual Athlon MP2400+, from a GeForce4 Ti4200 (NV25) to a GeForce 6800GT (NV40). Running Fedora Core 1 at the time, I was using a similar XFree86 release as RHEL/CentOS 3. That version includes only NV2x (GeForce3/4Ti) and early NV30 (GeForce FX5800) era support . I got the same results until I upgraded to the nVidia driver. Example 2: NV1x-era Windows XP with NV40 card Just this past April, I loaded _Windows_XP on a brand new Athlon64 3200+ system with a GeForce 6800GT (NV40) card. XP detected a GeForce compatible card and loaded the driver from when the NV1x series was current (GeForce2 era). I experienced the _exact_same_ issues in XP until I grabbed the latest WHQL driver for XP. Example 3: NV4x-era Xorg (FC3) with NV40 card Now on the _very_same_system_, I loaded Fedora Core 3 AMD64 which includes a newer Xorg release with not only NV40 (GeForce 6800), but even some NV41/43 (GeForce 6200/6600) and the latest NV45 (GeForce 6800, PCIe native). So I didn't have an issue even using the stock Xorg driver. So, are you sure you are using the nVidia driver _downloaded_ from nVidia? Or are you simply using the nVidia (nv) driver CentOS 3 is detecting when installed? I want to say, given your symptomps, you're using the later, which (even with update 4) is still a NV2x/NV30 driver, and not good for later NV31/34 (GeForce FX5200/5500) let alone NV4x (GeForce 6000 series) cards. And if you are using the nVidia driver, please note that you want to be running with the _latest_ 1.0-6xxx series driver (avoid the new 1.0-7xxx series driver until it matures). The 1.0-6xxx series driver is required for NV40 (GeForce 6800 series), and latter versions for NV41/43/NV45. From: Sukru TIKVES <sukru(a)cs.hacettepe.edu.tr> > As far as I know there are significant speed improvements between xorg > and XFree86 servers, but I may be wrong. The Xorg and XFree86 releases are pretty tandem. What most Red Hat users may assume is that because the last XFree86 in any Red Hat release (FC1/RHEL3) was from the NV2x/NV30 era, Xorg was added when the nv driver added NV3x/NV40 support (FC2+). XFree86 has matched those updates, but they just aren't in any newer releases in any Red Hat distro. > I guess CentOS is not suited for desktop machines, you may try your > chance with the upcoming FC4 release. (No, i am not talking about > workstations). Nope, that shouldn't be a factor. > [ Btw, my machine was Athlon XP 2800+ with GeForce FX 5200. Linux > install was smooth as in Windows 95 smooth on a P4. ] The GeForce FX5200 is the NV34 (or NV31?, I think that was the FX5900 or FX5700 -- the higher designation is not necessary faster, just newer, and can be actually an economy design). Anything Fedora Core 2+ (including CentOS 4) should support it, or possibly late Fedora Core 1 (and CentOS 3) with the final XFree86 updates. <TANGENT> BTW, I know people get into the big "proprietary" debate on nVidia. Just some facts to know of: 1) nVidia released the source code back in the XFree86 3.3.x days 2) Intel, Microsoft and other 3rd parties who owned IP in that release were not amused, and even though nVidia obfusciated the identifiers and structure of the C code, people started "cleaning it up" and exposing what Intel, Microsoft and others did not want. 3) XFree86 4.0 is MIT licensed introduced a modular design that supports closed source drivers, which nVidia and Matrox quickly took advanatge of. 4) It was at that time nVidia decided to "unify" it's entire NV0x on-ward driver model (Riva TNT and GeForce on-ward, but not earlier, pre-NV0x Riva/128). This has driven all their other decisions. 5) As part of the unified driver and its AGPgart and memory functions, nVidia was contractually forbidden to release it GPL -- by Intel itself. Intel considered AGP a "trade secret" implementation of PCI (AGP is _not_ a PCI Standards Group standard), hence why nVidia's AGP implementation and specifications were only in their closed video driver (not even the nForce -- it only worked for nForce+GeForce). The Intel AGPgart in the kernel is largely a clean-room design (and rather limited). NOTE: #5 is going away as with the release of PCI-Express by the PCI SG (an actual, open standard), Intel no longer considers AGP a "trade secret" and has freed up some of nVidia's obligations. There is already AGPgart code for nVidia chipsets in newer kernels, so you don't have to have a GeForce on a nForce chipset. 6) nVidia puts more developers on the MIT-licensed XFree86/Xorg drivers than anyone else. nVidia is completely open with its 2D drivers and specifications, including video-in/video-out and other scan converters. In a nutshell, the 2D MIT (nv) driver is very, very good as long as recent for the latest card support, because nVidia supports it. Of course, older releases with newer cards will always be a performance issue, and system-specific implementations (such as the Go Mobile series in notebooks) will definitely lack compatibility. The nVidia-licensed "nvidia" driver is a unifed 2D/3D with complete OpenGL-standards, ARB extensions as well as a few nVidia extension (this is far, far less common in the OpenGL world). The OpenGL on X11 (GLX) implementation is unmatched and without it, Linux would have not been so widely adopted as the primary visual workstation in the CAM and EDA world over the last 5-7 years (only as a back-end solution). If you just need 2D, you don't need the "nvidia" driver, not even for TV scan conversion. But the 3D version at least gives you an "open standard" platform, meaning you write applications to an "open standard" that does not require nVidia hardware. nVidia drivers continue to be trusted, based on open standard GLX performance and reliability. I see a few people complaining about the new 1.0-7xxx series of drivers, but that's not unexpected because it's a brand new driver release with OpenGL 2.0 and some major changes in how it handles VESA defaults (it's a bit more conservative, which is breaking things that used to be autodetected). 7) ATI tried to "go clean-room 3D" with its indirect support of UtahGLX and Direct Rendering Infrastructure (DRI) and releasing of the R100 series (Radeon 7500-8000) and R200 (Radeon 8500-9000) series. These 3D implementations were add-ons to the stock MIT drivers. Unfortunately, the lag time between release and lack of variant endless support (e.g., RV280/Radeon 9200, etc...) often caused major headaches and issues -- let alone performance sucked. Worse yet was the fact that nVidia had established itself as _the_ video card/ driver combo for production GLX. So as of the R300 (Radeon 9500+) series, ATI has _withheld_ 3D specifications. In reality, ATI has always withheld not only some 3D, but even 2D interfaces. This is largely because ATI has some really advanced 2D image encoding and other logic in their GPUs that are partially driven by software/algorithms licensed from a 3rd party (whereas nVidia largely off-chipsets much of its capabilities, or uses commodity approaches/algorithms that it has no issues documenting). Although some of the UtahGLX/DRI code still works for R300+ series cards, they are now severely lagging. Which means ATI now offers a "closed source" unified driver as well. 8) My final comments The reality is that video card products are _obsoleted_ within 12 months. GPUs double in performance 2-3 times faster than CPUs, which means driver development goes hand-in-hand with engineering development _pre-release_. In other words, we can dream of a day where there is no cutting-edge intellectual property in GPUs so vendors can release drivers as GPL, and not have to hack kernel interfaces because of 3rd party NDA obligations. But the idea that video card drivers can be community developed means that their drivers wouldn't be released until the GPU is already 1-2 generations behind. This is the reality of the product. Now, luckily, OpenGL/GLX are "open standards," and most vendor extensions are submitted and formalized by an architectural review board. This is very different than the DirectX world where "features" are often vendor-only extensions, and not always standardized in the API -- which is largely a wrapper to required, physical and often eccentric GPU functions until almost commodity 3-4+ generations behind and all video cards have them. This is why many applications in Windows require specific video cards, and not even after 2-3 generations the titles will run (if ever). So by supporting the OpenGL/GLX "open standards," even if you are running a proprietary, closed source driver to run your OpenGL/GLX application, it means you will be able to run it in the future on another video card. Even if the ARB extensions utilized and only supported by 1-2 "cutting edge" video cards (namely ATI and nVidia), they are typically supported by other cards within 1-2 generations by the very nature of how the OpenGL ARB works. </TANGENT> -- Bryan J. Smith mailto:b.j.smith@ieee.org

5 6

Re: [CentOS] Re: dragging windows leaves traces -- running older GeForce drivers on newer cards
by Bryan J. Smith <b.j.smith＠ieee.org> 06 Jun '05

06 Jun '05

From: "Hilliard, Jay" <Jay.Hilliard(a)disney.com> > We've had lots of trouble with nvidia's driver version 1.0-7174 which > was current last week. Yep, it's the first of a new series (7000 -- sounds like it'll support NV5x out-of-the-box), so that's expected. > Today there is a new driver (1.0-7664) which has fixed lots of issues > for us, specifically openGL applications and problems when dragging > windows around. Yep. They probably spent so much time testing OpenGL 2.0 that they missed some OpenGL 1.x compatibiltiy testing that end-user quickly found. At least they addressed it in short-order. -- Bryan P.S. I know there weren't any complaint about nVidia being "proprietary" on this list, but several people on the Fedora lists snickered about it not been open source as the problem. The reality is that if it was community developed, it would probably be available in another 6-12 months (much like ATI's prior UtahGLX/DRI endeavors were for newer cards). -- Bryan J. Smith mailto:b.j.smith@ieee.org

1 0

Re: [CentOS] Re: dragging windows leaves traces -- running older GeForce drivers on newer cards
by Bryan J. Smith <b.j.smith＠ieee.org> 06 Jun '05

06 Jun '05

From: Robin Mordasiewicz <robin(a)bullseye.tv> > Is it possible that some windowing libraries are responsible for slow > re-drawing. Re-sizing firefox seems slow, but resizing any native kde > application is quick. > resizing adobe acrobat reader is amazing slow, and moving the window > around the screen leaves a tracing streak. Ah-ha! Now that's some info that might provide a lead. Are you using KDE as your desktop? If so, it is Qt widget library based. Firefox is GTK+-based, and Acrobat is statically linked to Motif. If you can give more examples, maybe we can confirm this. Qt/KDE is built with some optional X11 (RENDER?) extensions which don't play nice with non-Qt apps (Xt, Xaw, GTK+, Motif, etc...). Didn't know this was the case on CentOS, or maybe it's some other configuration? I use GTK+/GNOME (gdm/nautilus/metacity session/file/window manager). -- Bryan J. Smith mailto:b.j.smith@ieee.org

1 0

Re: dragging windows leaves traces -- running older GeForce drivers on newer cards
by Bryan J. Smith <b.j.smith＠ieee.org> 06 Jun '05

06 Jun '05

From: Sukru TIKVES <sukru(a)cs.hacettepe.edu.tr> > Looks like I've complete misunderstood the issue. > Actually I've never thought it would still be problem. > I usually feel no difference between windows and linux. Correct. But in select operations, you are seeing this. That could be an application-specific setting, including the Window Manager which includes a number of applications. It's using a very slow, software-based set of buffers/overlays as you move the windows. Or it could be that application is attempting to use functions that implemented differently on newer GeForce cards than the driver which was designed for an older release. It's just hard to tell without seeing it or getting more, exact information. > Nevertheless as others mentioned it may be a window manager > (metacity is famed as "slow") or a Gtk2 issue (Gtk2 is getting > faster with each release aster 2.0). I've never had an issue myself, and I've been running with nVidia cards with both the MIT XFree86/Xorg "nv" and nVidia "nvidia" since first-generation NV0x Riva TNT/GeForce series on every GNOME 1.2+ through the latest 2.x revisions. I think either a setting is enabled, something Linux supports, but typically via software with a massive performance hit (much like what will happen with Avalon/WGF1.0 in the initial NT6.0 "Longhorn" release ;-), or you have a driver that is not current for your card and using functions that are deprecated or requires newer options. > Sorry for the irrelevant comments. I don't think they were irrelevant. At this point, the door is still very open on what the cause is. Again, it sounds like either an app setting or an app X11 call/ extension that the driver doesn't do well because it's dated. -- Bryan J. Smith mailto:b.j.smith@ieee.org

1 0

Re: [CentOS] Re: dragging windows leaves traces -- running older GeForce drivers on newer cards
by Bryan J. Smith <b.j.smith＠ieee.org> 06 Jun '05

06 Jun '05

From: Sukru TIKVES <sukru(a)cs.hacettepe.edu.tr> > NVidia has its own desktop utility for Windows. It works fine. It also > has a "keystone" function which allows you transform (stretch, skew, > resize, etc) your "entire desktop" freely. Actually, there is a GUI management program in the nVidia driver suite. Unfortunately, by the way X works, some of these things the Win32 GDI allows is not always possible in X, for a variety of reasons. > There is a new X extension called RENDER. It comes disabled with xorg > server, however once you enable the extension and install necessary user > level tools, you'll have hardware accelerated (ie: true) transparent > windows. Unfortunately, it's a hack. It uses unused video memory as a faster, local store than main memory, but still doesn't address the reality that you should be using your GPU functions instead of memory tricks. I.e., a window is simply a set of vectors and/or bitmaps on a plane. It doesn't matter if it looks like 3D to a user, or is perfectly perpendiular and appears to be 2D (but is still using 3D in the GPU, even if no Z buffer). >From simple concepts to the Xgl server to FreeDesktop Cario which brings full aliasing, vector graphics on planes as well as traditional bitmaps on planes, it is the X11/GLX's equivalent to QuartzExtreme. IMHO, I think Sun's got the right idea. Go 3D everything, solve the inconsistency, hacks upon hacks and overall inefficiency of layers upon layers with Looking Glass. What really sold me that Sun is "thinking ahead" is that even the input is 3D. The "technology demo" of Looking Glass doesn't do their innovative approach an ounce of justice. It's not eye candy, it's efficiency. > See my previous post. I've tried the setup on a gentoo installation > and it works very well. On nVidia cards with the nVidia driver, but that's about all right now. It largely has to do with the fact that nVidia driver has that proprietary memory management module in the kernel that is tied to the X11/GLX driver/libraries. nVidia knows what it's doing, because they solved the problem 5+ years ago. -- Bryan J. Smith mailto:b.j.smith@ieee.org

1 0

When is the next CentOS release?
by Sukru TIKVES 06 Jun '05

06 Jun '05

Hello all, There are significant kernel defects in CentOS 4 which affects our servers. Our most serious bug is mentioned in http://bugs.centos.org/view.php?id=870 however it says: "Some of these kernel issues are known to have been fixed upstream and should make the 4.1 Release ( due soon now )" How soon is the release of the CentOS 4.1 version? -- sukru

4 3

FrontPage Extensions
by Bob Pierce 06 Jun '05

06 Jun '05

Has anyone had success getting FrontPage extensions to work with Centos-4. I've been trying to set up a virtual hosting server with front page support. I found this page to help getting around the selinux problems: http://www.rtr.com/fp2002disc/_disc2/00000be5.htm but each problem I overcome I keep hitting another... Just wondering if anyone's had success with this on Centos-4 or if you've found any good tips to share. Thanks for your help, Bob Pierce.

4 3

Re: S/N on this list is plummeting...
by Bryan J. Smith <b.j.smith＠ieee.org> 06 Jun '05

06 Jun '05

Ugo Bellavance wrote: > The thing is that sometimes threads start as technical and end up by > opninated, endless discussions. It depends. Sometimes people just differ on what context they are using, so they will differ by nature. I think the comments on Red Hat in general just need to die though. It's clear there are some people just don't like Red Hat, and no amount of discussion is going to change that. From: Chris Mauritz <chrism(a)imntv.com> > Exactly! I guess I can just set up some custom filters for the "biggest > offender(s)," I'm definitely in that group. ;-> > but it's a shame that it has to come to that. Looking at the sheer > number of posts in such a short timeframe from some of these > folks leaves me envious of their amount of free time (or perhaps just a > superhuman typing ability). 8-) I am a writer by trade. I can crank out a productive 45 words/minute. 5-15 words/minute effective for publication (i.e., including draft, review, etc...). -- Bryan J. Smith mailto:b.j.smith@ieee.org

7 11

Re: Is it possible to start Installer from WindowsXP?
by Bryan J. Smith <b.j.smith＠ieee.org> 06 Jun '05

06 Jun '05

From: Sukru TIKVES <sukru(a)cs.hacettepe.edu.tr> > Of course, I know that! > That's why I said "Make a FAT partition" and "Boot DOS (probably > via a floppy)". > BTW: You can use a Windows ME/98/95 installation CD to boot into DOS. Yes, but unfortunately this typically messes up the BIOS disk assignments, and the DOS redirector (normally used for networking, but also CD) also plays games with memory. What you will need to do is boot directly to an actual DOS partition with a DOS installation ("Chicago" DOS7 aka Win9x/Me will work). You might be able to get away with booting as long as the LOADLIN binary and kernel/initrd are not on the CD so the redirectory doesn't need to be loaded. In a nutshell, it's always a PITA, regardless. -- Bryan J. Smith mailto:b.j.smith@ieee.org

1 0

Is it possible to start Installer from WindowsXP?
by Takashi Nakamura 06 Jun '05

06 Jun '05

Hi, My computer has only primaly IDE but not secondaly. I connected a hard drive as Master and DVD-R as Slave on IDE. BIOS doesn't recognize IDE devices, but Windows XP that boots up from SATA hard drive can recognize DVD-R and can read files on the CentOS Installer Disk 1 of 4. Is it possible to start Centos4 Installer from WindowsXP? If possible, which file on Disk 1 should I click? Even if they are possible, can CentOS boot up from IDE Master hard drive? I think probably not. Thanks in advance, Takashi Nakamura

3 3

CentOS-announce Digest, Vol 4, Issue 3
by centos-announce-request＠centos.org 06 Jun '05

06 Jun '05

Send CentOS-announce mailing list submissions to centos-announce(a)centos.org To subscribe or unsubscribe via the World Wide Web, visit http://lists.centos.org/mailman/listinfo/centos-announce or, via email, send a message with subject or body 'help' to centos-announce-request(a)centos.org You can reach the person managing the list at centos-announce-owner(a)centos.org When replying, please edit your Subject line so it is more specific than "Re: Contents of CentOS-announce digest..." Today's Topics: 1. CESA-2005:416-01: Low CentOS 2 i386 kdbg security update (John Newbigin) 2. CESA-2005:480-01: Moderate CentOS 2 i386 ImageMagick security update (John Newbigin) 3. CESA-2005:481-01: Low CentOS 2 i386 openssh security update (John Newbigin) ---------------------------------------------------------------------- Message: 1 Date: Mon, 06 Jun 2005 08:52:58 +1000 From: John Newbigin <jnewbigin(a)ict.swin.edu.au> Subject: [CentOS-announce] CESA-2005:416-01: Low CentOS 2 i386 kdbg security update To: centos-announce(a)centos.org Message-ID: <42A3824A.5060504(a)ict.swin.edu.au> Content-Type: text/plain; charset=us-ascii; format=flowed The following errata for CentOS-2 have been built and uploaded to the centos mirror: RHSA-2005:416-01 Low: kdbg security update Files available: kdbg-1.2.1-7.i386.rpm More details are available from the RedHat web site at https://rhn.redhat.com/errata/rh21as-errata.html The easy way to make sure you are up to date with all the latest patches is to run: # yum update -- John Newbigin Computer Systems Officer Faculty of Information and Communication Technologies Swinburne University of Technology Melbourne, Australia http://www.ict.swin.edu.au/staff/jnewbigin ------------------------------ Message: 2 Date: Mon, 06 Jun 2005 08:53:01 +1000 From: John Newbigin <jnewbigin(a)ict.swin.edu.au> Subject: [CentOS-announce] CESA-2005:480-01: Moderate CentOS 2 i386 ImageMagick security update To: centos-announce(a)centos.org Message-ID: <42A3824D.8050102(a)ict.swin.edu.au> Content-Type: text/plain; charset=us-ascii; format=flowed The following errata for CentOS-2 have been built and uploaded to the centos mirror: RHSA-2005:480-01 Moderate: ImageMagick security update Files available: ImageMagick-5.3.8-11.c2.1.i386.rpm ImageMagick-c++-5.3.8-11.c2.1.i386.rpm ImageMagick-c++-devel-5.3.8-11.c2.1.i386.rpm ImageMagick-devel-5.3.8-11.c2.1.i386.rpm ImageMagick-perl-5.3.8-11.c2.1.i386.rpm More details are available from the RedHat web site at https://rhn.redhat.com/errata/rh21as-errata.html The easy way to make sure you are up to date with all the latest patches is to run: # yum update -- John Newbigin Computer Systems Officer Faculty of Information and Communication Technologies Swinburne University of Technology Melbourne, Australia http://www.ict.swin.edu.au/staff/jnewbigin ------------------------------ Message: 3 Date: Mon, 06 Jun 2005 08:53:04 +1000 From: John Newbigin <jnewbigin(a)ict.swin.edu.au> Subject: [CentOS-announce] CESA-2005:481-01: Low CentOS 2 i386 openssh security update To: centos-announce(a)centos.org Message-ID: <42A38250.9010705(a)ict.swin.edu.au> Content-Type: text/plain; charset=us-ascii; format=flowed The following errata for CentOS-2 have been built and uploaded to the centos mirror: RHSA-2005:481-01 Low: openssh security update Files available: openssh-3.1p1-18.i386.rpm openssh-askpass-3.1p1-18.i386.rpm openssh-askpass-gnome-3.1p1-18.i386.rpm openssh-clients-3.1p1-18.i386.rpm openssh-server-3.1p1-18.i386.rpm More details are available from the RedHat web site at https://rhn.redhat.com/errata/rh21as-errata.html The easy way to make sure you are up to date with all the latest patches is to run: # yum update -- John Newbigin Computer Systems Officer Faculty of Information and Communication Technologies Swinburne University of Technology Melbourne, Australia http://www.ict.swin.edu.au/staff/jnewbigin ------------------------------ _______________________________________________ CentOS-announce mailing list CentOS-announce(a)centos.org http://lists.centos.org/mailman/listinfo/centos-announce End of CentOS-announce Digest, Vol 4, Issue 3 *********************************************

1 0

Oracle Installation +CentOS4
by Fatih Ölçer 06 Jun '05

06 Jun '05

Hi, I've tried to install Oracle Database 10g on CentOS but installation doesnt start saying op. system should be RHEL 3, SUSE 9.0... (or upper) etc. does anyone knows how to install Oracle Database 10g on CentOS. Should I modify runinstall script? Thanks.

3 2

Backspace key and gnome-terminal
by Oliver Schulze L. 05 Jun '05

05 Jun '05

Hi, My Delete key does not work anymore in gnome-terminal. I have a keyboardless server, so I think this may be the problem. When I press the Delete key, I get the ~ caracter printed. I deleted all the ~/.g* files, in order to have a clean and default gnome installation but still, no luck. This is my locale: LANG=en_US.ISO-8859-1 LC_CTYPE="en_US.ISO-8859-1" LC_NUMERIC="en_US.ISO-8859-1" LC_TIME="en_US.ISO-8859-1" LC_COLLATE="en_US.ISO-8859-1" LC_MONETARY="en_US.ISO-8859-1" LC_MESSAGES="en_US.ISO-8859-1" LC_PAPER="en_US.ISO-8859-1" LC_NAME="en_US.ISO-8859-1" LC_ADDRESS="en_US.ISO-8859-1" LC_TELEPHONE="en_US.ISO-8859-1" LC_MEASUREMENT="en_US.ISO-8859-1" LC_IDENTIFICATION="en_US.ISO-8859-1" LC_ALL= When I start gnome, I get a warning about not having the XKB extension loaded. Any gelp? Thanks Oliver -- Oliver Schulze L. <oliver(a)samera.com.py>

1 0

DVD writer problem
by Mark Jarvis 05 Jun '05

05 Jun '05

I have an NEC DVD-RW ND2500A (that's not the label I bought it under, but that's what it really is.) It's supposed to read/write CDs, DVDs, DVD+Rs, and DVD-Rs. I'm basically dual booting between XP and CentOS 4. Following the recipe in "The Linux Cookbook", I tried to copy a DVD using: dd if=/dev/hdd of=/tmp/diskfile.iso then cdrecord dev=/dev/hdd fs=32m -sao -v -eject -dummy /tmp/diskfile.iso to test and cdrecord dev=/dev/hdd fs=32m -sao -v -eject /tmp/diskfile.iso to actually record. I had a lot of trouble getting a good read of the whole original DVD--it kept getting an I/O error at 94-95% of the way through. After multiple cleanings and 20-40 attempts, I finally got the whole thing copied with no reported errors. It then took me a couple of tries to get the cdrecord options right and to run as root to get the requisite priority. The "-dummy" run ran with no errors that I noticed--there was an LOT of output--and "-dummy" didn't work. It wrote to the blank DVD anyway. I then did a run without the "-dummy". Although the original will play on both my PC and the DVD player hooked to the TV, the two copied DVDs (DVD+Rs) are coasters. DURN! Any solutions, suggestions, pointers to articles I should read, etc. will be appreciated. BTW, I've never quite gotten around to playing DVDs or my music files (mp3 and Real Player .rmj files) from Linux. I'd appreciate being pointed in the direction to start. -mj-

4 5

Vote For CentOS :)
by Johnny Hughes 05 Jun '05

05 Jun '05

The "2005 Readers' Choice Awards" voting is in progress at Linux Journal, and CentOS is Nominated in the "distribution:" category. So ... all you guys and gals who think CentOS is the best thing ever, get over to: http://www.linuxjournal.com/article/8266 and vote for CentOS :) ------------------------------- Also, since write ins are allowed on the e-mailed ballot, if you don't have a favorite for the "web hosting service:" category, might I suggest you pick your favorite from the hosting companies on our website who donate servers to the CentOS project: http://www.centos.org/mirrors/ (one hosting company who is making a server donation now, but it isn't quite setup yet ... and isn't on the donor list just yet is: http://www.eurovps.com/ ) ------------------------------- This is the first round of voting, which ends June 27th and then there will be a final ballot. Make sure to send in the e-mail ballot in plain text ... and not HTML. Thanks, Johnny Hughes

18 68

intel pro 2200 dsl modem
by Robert Hanson 04 Jun '05

04 Jun '05

greetings does or has anyone know(n) of a solution for linux to interface with the intel pro 2200 internal dsl modem please? thanks - rh

2 4

LVM problems [ was: bad blocks ... ]
by Collins Richey 04 Jun '05

04 Jun '05

On 6/4/05, John Hinton <webmaster(a)ew3d.com> wrote: > > Otherwise, I did have troubles with a LVM install and decided to never > try that one again... at least for a long time. > Now you have my curiosity aroused. Was this with CentOS3 or 4, ie LVM or LVM2? What problems? We use LVM on all our RH9 setups at work. No problems at all. OTOH, LVM2 is a work in progress that RH chose to pass along. I keep hearing rumors that the support may be more complete in Update 1 whenever that comes along. -- Collins Head teachers of the world unite: you have nothing to lose but the Start button.

2 1

Is mod_auth_ldap in a Centos 4 repo anywhere?
by Ted Kaczmarek 04 Jun '05

04 Jun '05

Is their a repo with mod_auth_ldap ? Regards, Ted

3 2

bad blocks showing up
by Jerry Geis 04 Jun '05

04 Jun '05

I am getting a few bad sector messages in my /var/log/messages. I have read where an "fsck -c -c /dev/hda" may be what I need. Before I go doing such things I am looking for confirmation that that is what I should do. Anyone please comment on how to tell linux to not use sectors in my disk. This is stright IDE no raid not nothing at this point /dev/hda is all. Thanks, jerry

6 5

rpmbuild question about debugfiles.list
by Robin Mordasiewicz 04 Jun '05

04 Jun '05

I am trying to rebuild multimedia rpms for CentOs3.4 I have downloaded povray from thac's site. Near the packaging prt of the build I get an error.... Processing files: povray-debuginfo-3.1g-19.el3 error: Could not open %files file /usr/src/redhat/BUILD/povray31/debugfiles.list: No such file or directory RPM build errors: Could not open %files file /usr/src/redhat/BUILD/povray31/debugfiles.list: No such file or directory I have rebuilt many other rpms without this error. Can someone explain what is going on here.

1 0

Re: [CentOS] Thanks :)
by Bryan J. Smith <b.j.smith＠ieee.org> 04 Jun '05

04 Jun '05

From: Ajay <ajay(a)unisoftindia.net> > and hope that Redhat accepts the existance of this project and finds a > way to utilize it for good of both. You were doing fine up until this point. I think the speculation on Red Hat needs to stop, and I don't know why you had to interject this as well. From what I've seen, I don't think Red Hat is in any legal position to accept the existance of any project, and I no longer want to go down the road of where Red Hat's "hands are tied" on the nagging trademark history. So let's just leave it be. -- Bryan J. Smith mailto:b.j.smith@ieee.org

3 2

Re: [CentOS] Vote For CentOS :)
by me＠prestoncrawford.com 04 Jun '05

04 Jun '05

-------- Original Message -------- > From: Alex White <prata(a)kuei-jin.org> > Sent: Thursday, June 02, 2005 5:45 PM > To: CentOS mailing list <centos(a)centos.org> > Subject: Re: [CentOS] Vote For CentOS :) > > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Johnny Hughes wrote: > | > | No ... not at all. RedHat does not spend any time making sure > that the > | people who use CentOS can get free updates. > | > | They did not put together a world wide mirror network for free that > | everyone can use to get free enterprise level software. > | > | I am sick and tired of people minimizing the work that the CentOS > | Developement Team does. > | > | If you don't want to use CentOS, then don't use it ... but it is not > | easy to maintain a distro based on RHEL ... (or anything else for that > | matter). > | > | So, no, if you like and use CentOS, you should vote for CentOS. > | > | A couple hundred thousand people use it ... and hopefully they > | appreciate the amount of work it takes to maintain. > > Johnny, > > I certainly appreciate all of the work that has gone into CentOS. > I've not done that ballot just yet, but I will certainly be casting > my vote for CentOS. > > Thanks Johnny for the work. I appreciate it as well. In fact I plan on donating one of these days when I get the money. And I'm definitely recommending it to all my friends/family who use Linux. Yes, Red Hat creates RHEL. But by that logic shouldn't the credit be given to each contributor to each individual project that goes into a distro? Red Hat doesn't create mySQL, do they? Or Apache? How about Firefox? Is that Red Hat's baby? Point is, if you keep on saying "give credit where credit is due" you have to pass it down the line to the individual developers, testers, bug-fixers, etc. for the thousands of projects that go into RHEL. Not to mention all the brave Fedora users. So in the end, all things being equal and considering the above, my vote goes for CentOS. Preston

5 5

Re: [CentOS] Vote For CentOS :)
by Bryan J. Smith <b.j.smith＠ieee.org> 04 Jun '05

04 Jun '05

From: Robert <kerplop(a)sbcglobal.net> > You noticed that statement, too, huh? Yes. That's why I now sound like a broken record: We _all_ stand on the shoulders of others No one should be considered more important than anyone else, at least when it comes to GPL software and redistribution. The great thing about any GPL project is that if you don't like them, the project, etc... you can choose another. Unfortunately, too many people think you must comment negatively on the original, or the fork. There is all this "versus" and "absolutes" non-sense that drives me wild. > That begs the questions: > Why am I here then? > Why are you (not you, BS) here then? I don't claim to be innocent. But other than stupidly engaging on the "GPL license/trademarks" stuff, I maintain a 100% _technical_ focus in my commentary. I could care less about credit, brandnames, marketing, etc... > Why are we running CentOS? From dictionary.com: 3. Worth > in usefulness or importance to the possessor; utility or merit: > the value of an education. I think people get confused on "value." Yes, if RHEL wasn't available in SRPM form, or didn't exist, then CentOS probably wouldn't exist (at least as we know it). And sure, without CentOS, RHEL would still go on. But that doesn't mean CentOS doesn't have value. -- Bryan J. Smith mailto:b.j.smith@ieee.org

3 2

installation count?
by Todd Blake 04 Jun '05

04 Jun '05

Out of curiousity...I wonder if anyone has an idea of how many centos installations are out there?

5 7

CentOS/RHEL versioning scheme?
by Morten Kjeldgaard 04 Jun '05

04 Jun '05

I am a bit puzzled at the versioning scheme of the RedHat clone family. RedHat seems to use integer 4, Tao and Centos does the same. If you do rpm -q --qf '%{version}\n' -f /etc/redhat-release you get '4'. However, Scientific Linux uses 4.0, and that seems to me to be a more logical choise, since presumable there are going to be versions 4.1, 4.2, etc. Is there any good reason for the version of package centos-release NOT to be 4.0? It has a significance in automated scripts trying to work out which distribution and version you are running, and it seems silly to treat the different RHEL4 clones differently. It would make sense for CentOS to coordinate/agree_upon a common scheme among the different distributions. Cheers, Morten PS: Perhaps someone can provide info about other RHEL clone distros! -- Morten Kjeldgaard, Asc. professor, Ph.D. Department of Molecular Biology, Aarhus University Gustav Wieds Vej 10 C, DK-8000 Aarhus C, Denmark Lab +45 89425026 * Mobile +45 51860147 * Fax +45 86123178 Home +45 86188180 * ICQ 27224900 * http://www.bioxray.dk/~mok

4 5

Re: [CentOS] Attempting custom CentOS4 network install
by Bryan J. Smith <b.j.smith＠ieee.org> 04 Jun '05

04 Jun '05

From: Ajay <ajay(a)unisoftindia.net> > A new hdlist and hdlist2 is sucessfully regenerated in CentOS/base dir > with a small reduction in file size of hdlist, but the network install fails > at the end of "Preparing RPM transaction" screen with the message > "install exited abnormally" and shuts down. > Am I missing something? I've had this happen when there is a conflict between some core packages. It's a PITA to resolve because, as you found out first-hand, you don't know it until you're half-way through an install. My approach has been to only add packages I know are safe and won't conflict with core FC/RHEL packages, and then install any additional from either an internal APT-RPM or YUM-RPM repository. I typically do the latter to fetch system updates anyway, including creation of RPMs for key configuration files (and updates). E.g., "CONFIG_BS-2.1.noarch.rpm" (the package name conviniently matching my initials if anyone hasn't noticed that yet ;-). From: Sean O'Connell <oconnell(a)soe.ucsd.edu> > Ajay- > It would be far simpler to setup a yum repository for the additional > packages, create a custom yum.conf (or repo file for the repo), and then > install that yum.conf or repo file in %post and then follow that up with > a > yum -t -y install foo bar baz I remember Seth talking awhile back on a Fedora list about the next generation installer (Anaconda-based?) being able to pull from YUM-RPM repositories at install-time for network installs. Has anyone been tracking that development? -- Bryan J. Smith mailto:b.j.smith@ieee.org

4 4

multimedia kernel
by Robin Mordasiewicz 03 Jun '05

03 Jun '05

Is anyone using any type of multimedia patched kernels such as the one available at planet CCRMA. Or does anyone know where I can dl multimedia patched kernel rpm for centos.

3 6

Re: [CentOS] bad blocks showing up
by Bryan J. Smith <b.j.smith＠ieee.org> 03 Jun '05

03 Jun '05

From: Ted Kaczmarek <tedkaz(a)optonline.net> > I have seen what I consider an exorbitant amount of ide drive failure > lately, granted most of these are lower grade machines, but the > percentage of drives failing seems to be much higher than I recall. That's because ATA drives are spinning higher and higher, and people aren't cooling nearly as good as they used to. The typical, "commodity" ATA drive is typically rated for only 50,000 restarts with 8 hours of operation between, resulting a MTBF of 400,000 hours. And that assumes the following "ideal" conditions: - Ambient drive temperature never exceeds 40 degrees C - Drive is not powered 24x7 (most vendors recommend no more than 14x5 on their "commodity" models) - Drive is spun up at least once every few weeks** [ **NOTE: This is a major reason why you should _never_ use ATA drives a backup media if they are not going to be powered-on regularly (like in an on-line/near-line backup system). ] Now some vendors test their "commodity" drive lines and reserve those that test to higher tolerances for "Enterprise" labeling, possibly for the same ATA, or maybe a server SATA or SCSI drive. Other vendors use completely different lines for "enterprise" drives, using better lubricants, filters, mechanics, etc... > Is their possibly something with ext3 that stresses drives more? > Maybe the location of the journaling info? Naah, that would affect FAT systems and others too. The main problem is the density, spindle, heat and "ultra-thin manufacturing margin" of the drives these days. Especially given the increased spindles, the "cheap drives" just don't last under the constant operation and heat at 7200rpm. The "Big 3" manufacturers are Hitachi, Maxtor and Segate. IBM, Western Digital and other vendors no longer manufacture, and haven't for awhile. Hitachi and Seagate have split-lines, commodity and enterprise. Interface doesn't have anything to do with it, and even Hitachi and Seagate have SATA and SCSI models that use the exact same lines as their most commodity ATA ones. Of course, they also offer "Enterprise" ATA, SATA and SCSI models too, that are rated for much higher tolerances. Maxtor actually has a single commodity approach. Then they test drives against tolerances and rate them for MTBF and warranty based on that testing. I.e., all drives pretty much come off the same line, but the "Enterprise" drives are the ones that test high, while the "commodity" drives are everything left over. I'm partial to Seagate, despite the extra 10-20% cost on the commodity models, because they are giving 3-5 warranties as standard. They seem to do a little better on standard components, but that's just my opinion. -- Bryan J. Smith mailto:b.j.smith@ieee.org

1 0

Re: [CentOS] Re: Separate list for political/philosophical discussion
by Bryan J. Smith <b.j.smith＠ieee.org> 03 Jun '05

03 Jun '05

From: Ugo Bellavance <ugob(a)camo-route.com> > What does that mean... I can't really connect it to my original post. Sorry, I meant "list" instead of "last." As always, I believe in a meritocracy. And that means those who are on the project have final say. And those who don't should respect their wishes. With that said, I am not on the project, so I would respect what the project maintainers say. -- Bryan J. Smith mailto:b.j.smith@ieee.org

3 2

opensource backup software suggestions please
by Robin Mordasiewicz 03 Jun '05

03 Jun '05

What are the users here opinions on enteriprise grade opensource software for doing backups, as an alternative to bakbone or veritas. We just have basic requirements, i.e backing up nfs servers. --

8 9

Re: [CentOS] Re: opensource backup software suggestions please -- VTL option ...
by Bryan J. Smith <b.j.smith＠ieee.org> 03 Jun '05

03 Jun '05

From: Ugo Bellavance <ugob(a)camo-route.com> > I'm using Thunderbird with nntp access on gmane and it works perfectly, > maybe you should try this out Brian. I use Evolution at home and it has no issues tracking Message Id. > I used to break threads with my other e-mail clients. Unfortunately, I have to use a web-based client at work. -- Bryan J. Smith mailto:b.j.smith@ieee.org

1 0

Re: opensource backup software suggestions please -- VTL option ...
by Bryan J. Smith <b.j.smith＠ieee.org> 03 Jun '05

03 Jun '05

From: Jim Bartus <jbartus(a)advance.net> > Bryan, > Something about your replies causes them to always be new threads rather > than inline with the rest of the conversation. > Is this just me (thunderbird 1.0.2 on FC3) or does everyone get this? I don't think my webmail reader is honoring the Message ID in the header. And I know my Treo doesn't. I'm working on switching to a new webmail interface. I hate it because I have to manually bottom-post. Unfortunately, GMail doesn't honor Message ID well either, so it's no better. I might switch to Yahoo. -- Bryan J. Smith mailto:b.j.smith@ieee.org

3 2

Topics in the CentOS-Announce list
by Johnny Hughes 03 Jun '05

03 Jun '05

I have been working with Joe Pruett (thanks Joe), to get the Topics feature for the mail list working. I think we now have the topics working fairly well. So if you only care about CentOS-3 i386 announcements (for example), you can pick that specific topic. There is also an option to pick the topics you care about, and all items that do not match any topics. (So that if we make an announcement that is not tied to one of those categories, you can get it too). If you are subscribed to the announce list and want to pick specific topics, go to: http://lists.centos.org/mailman/options/centos-announce Login with your e-mail address and password and choose the topics you want to receive. -- Johnny Hughes <http://www.HughesJR.com/>

1 0

Re: Vote For CentOS :) -- standing on the shoulders of each other
by Bryan J. Smith <b.j.smith＠ieee.org> 03 Jun '05

03 Jun '05

On Fri, 2005-06-03 at 09:36, Simon Perreault wrote: > i'm not sure you've considered your position thoroughly. > I am, and I will say it again just for the shock value: > Red Hat is the main developer of CentOS. On Fri, 3 Jun 2005, Les Mikesell wrote: > I'd put it this way instead: Red Hat is responsible for any > difficulty in creating the CentOS distribution, while sharing > the same upstream developers as all other Linux distributions. From: Maciej ?enczykowski <maze(a)cela.pl> > BULLSHIT And so it begins. I think we've all had our say at this point. I can understand (although not agree) with Les' viewpoint. And I'm sure Simon feels like I do, Red Hat doesn't get due credit. But this is why I dislike these "distro contests" in the first place. The are entirely subjective and full of "absolutes" and "versus." Reality is that there is much so much overlap. *** Because we all stand on the shoulders of each other! *** Vote CentOS if you like it best. Vote RHEL if you like it best. Do it on your own opinion for your own reasons. I laid out what I thought mine were before. At least one person liked it. I try not to talk in "absolutes." And try to understand why RHEL and CentOS exists, let alone the Fedora Core and Red Hat Linux lineages. Apparently people want to debate this from a marketing/branding, and that's _always_ going to be "subjective." There are those of us who aren't upset with Red Hat, and deploy _both_ RHEL and CentOS for various reasons. And there are those others of us who are upset with Red Hat, and believe CentOS should not have to exist, that Red Hat should be providing it. And I'm sure there are others who wonder why people don't just run Fedora Core, or another distro for that matter. And yet other viewpoints different from those. It's all GPL, so lets just appreciate each other. This is what I meant by people focusing more on ripping companies that produce GPL incompatible software, like IBM, Sun (sans OpenOffice.org) etc... <Flammatory=on> Red Hat v. CentOS, etc... -- it's like being a Christian watching Cathlotics and Protestants argue their religious viewpoints in Northern Ireland. Because in the end, it's really about focusing on differences than similarities -- and it doesn't matter how close those similarities are, people can find differences if they want. Sorry for this political analogy, but it pretty much fits well. At least from my ignorant American viewpoint. </Flammatory> -- Bryan J. Smith mailto:b.j.smith@ieee.org

1 0

Re: [CentOS] Vote For CentOS :)
by Bryan J. Smith <b.j.smith＠ieee.org> 03 Jun '05

03 Jun '05

From: Simon Perreault <nomis80(a)lqt.ca> > It does not give added value over RHEL in the distro itself. Be careful with that statement. ;-> Also remember that CentOS is more than just the RHEL distro. -- Bryan J. Smith mailto:b.j.smith@ieee.org

2 1

Triple /proc/net/stat/ip_conntrack files
by Maciej Żenczykowski 03 Jun '05

03 Jun '05

[root@tcs stat]# uname -a Linux tcs 2.6.9-5.0.5.ELsmp #1 SMP Wed Apr 20 00:16:40 BST 2005 i686 i686 i386 GNU/Linux [root@tcs stat]# pwd /proc/net/stat [root@tcs stat]# ls -al total 0 dr-xr-xr-x 2 root root 0 Jun 3 18:51 . dr-xr-xr-x 5 root root 0 May 31 23:12 .. -r--r--r-- 1 root root 0 Jun 3 18:51 arp_cache -r--r--r-- 1 root root 0 Jun 3 18:51 ip_conntrack -r--r--r-- 1 root root 0 Jun 3 18:51 ip_conntrack -r--r--r-- 1 root root 0 Jun 3 18:51 ip_conntrack -r--r--r-- 1 root root 0 Jun 3 18:51 ndisc_cache -r--r--r-- 1 root root 0 Jun 3 18:51 rt_cache Anyone getting the above triplication? Cheers, MaZe

1 0

Re: [CentOS] Re: opensource backup software suggestions please -- VTL option ...
by Bryan J. Smith <b.j.smith＠ieee.org> 03 Jun '05

03 Jun '05

From: Karanbir Singh <Mail-Lists(a)karan.org> > Which is a good thing really, since you can snip off the irrelevant bits > on your way down. Yes, I do it. But in my case, I've got to manually cut'n paste and add the greater-than signs. It gets tedious (especially on my Treo). But I do it because I believe in bottom-posting. > Gmail works fine. It sets the In-Reply-To: header. I am sure your move > to a more not-as-much-brain-dead client would be appreciated. Yeah, at least GMail would be better than this account. -- Bryan J. Smith mailto:b.j.smith@ieee.org

1 0

Re: opensource backup software suggestions please -- VTL option ...
by Bryan J. Smith <b.j.smith＠ieee.org> 03 Jun '05

03 Jun '05

From: Chris Mauritz <chrism(a)imntv.com> > I've found that backing up to another disk to be both faster and more > practical. We've got a number of multi-terabyte RAID arrays to store > uncompressed video. Backing all that data up to tape just isn't a > tenable solution. So we've bitten the bullet and built an extra larger > multi-terabyte array to do online backups. Periodically we do snapshots > of the data on that backup array (like you do on a netapp filer). That > system works well for us. It wouldn't cover us if the building burned > down so the next step is an offsite mirrored array. If you have the dough, Virtual Tape Libraries (VTLs) are the ultimate in flexiblity. Although there are many solutions, nearly all of them are the Red Hat Linux 7.3 + FalconStor Software combination. Although EMC, IBM and others have traditionally use standard disk arrays as they use for other solutions (with the software slapped on), Copan Systems' hardware solution is designed specifically for VTL and VTL-only. The density, power/thermal, cost and other considerations are rather interesting. -- Bryan J. Smith mailto:b.j.smith@ieee.org

3 2

Re: [CentOS] Separate list for political/philosophical discussion
by Bryan J. Smith <b.j.smith＠ieee.org> 03 Jun '05

03 Jun '05

From: Ugo Bellavance <ugob(a)camo-route.com> > Hi, > How about creating another mailing list to let people vent themselves? > It would make it easier for everyone, since technical/troubleshooting > issues and opinions target different people. That's how most projects handle it. Because invariably, people will start breaking things down into marketing, legal and other meta-discussions. And I'm just the type of person that deals with them on a daily basis, so I have my viewpoints. Rule of thumb on what is on-topic for that last should default to the project maintainers, and we should respect their viewpoints. -- Bryan J. Smith mailto:b.j.smith@ieee.org

2 1

Re: CentOS/RHEL versioning scheme? -- adversely affecting dist-tag
by Bryan J. Smith <b.j.smith＠ieee.org> 03 Jun '05

03 Jun '05

From: Morten Kjeldgaard <mok(a)bioxray.dk> > I am a bit puzzled at the versioning scheme of the RedHat clone family. > RedHat seems to use integer 4, Tao and Centos does the same. If you do > rpm -q --qf '%{version}\n' -f /etc/redhat-release > you get '4'. > However, Scientific Linux uses 4.0, and that seems to me to be a more > logical choise, since presumable there are going to be versions 4.1, > 4.2, etc. Unfortunately, the "release" file impacts build environments. E.g., dist-tag. So it shouldn't be done on a dime. Even Red Hat has really pulled some doozers -- like getting rid of revisions, restarting the numbering on Fedora Core to 1 which threw a wrench into the dist-tag through Red Hat Linux 9, etc... [ I can only assume there was a more important, legal reason, like not showing product lineage to protect Red Hat(R)? ] -- Bryan J. Smith mailto:b.j.smith@ieee.org

1 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Discuss