From: baldyeti <e_...@ho...> - 2007-04-13 17:52:12
|
Hello, I currently have colinux 0.6.3 setup to boot my HD-installation of debian sarge. What I really like and (as far as i can tell) only colinux can offer is the ability to run the very same linux i can start via dual-boot, not just an image. That's neat. Now sarge is aging, and i'd like to either upgrade or re-install the recently released etch version. I am wondering, though, if this has a chance to keep on working with the older kernel official colinux uses? Would a snapshot of 0.7.1 or even 0.8 be required to run an etch install (particularly wrt udev). |
From: Henry N. <Hen...@Ar...> - 2007-04-16 18:52:43
|
baldyeti wrote: > Hello, I currently have colinux 0.6.3 setup to boot > my HD-installation of debian sarge. What I really > like and (as far as i can tell) only colinux can offer > is the ability to run the very same linux i can start > via dual-boot, not just an image. That's neat. > Now sarge is aging, and i'd like to either upgrade or > re-install the recently released etch version. > I am wondering, though, if this has a chance to keep > on working with the older kernel official colinux uses? > Would a snapshot of 0.7.1 or even 0.8 be required to > run an etch install (particularly wrt udev). I'm not known about Debian versions and kernel depencies. From my installations I know coLinux 0.6.4, 0.7.1 and 0.8.x runs on a Debian Woody. The coLinux limit is only, you should have the same binaries for Windows, the Linux kernel and the modules. Your system binaries (under /bin, /sbin and so) can be different. Unpack the Modules vmlinux-modules.tar.gz from new colinux under /lib/modules/2.6.12-co-0.7.1, this is the only upgrade you need in your dual boot. CoLinux versions 0.6.4, 0.7.1 and 0.8.0 runs also on my SuSE 9.0 dual boot with an original 2.4.21 kernel. From my SuSE I know, that you can run every coLinux on an older distry. For very old distries with kernel 2.4.x you need to upgrade only the module-init-tools with fallback for 2.4-kernels. You can not run a new coLinux Windows (*.exe) with an older coLinux kernel from an other build. Eatch coLinux release Version supports only exactly one kernel version. -- Henry |
From: baldyeti <e_...@ho...> - 2007-04-17 09:30:37
|
Thanks Henri. But Woody and 2.4.x kernels are definitely even older than the sarge I want to replace! Etch comes with a 2.6.18 kernel and uses udev - that was my primary concern. So I if read you correctly, colinux 0.7.0 will come with a 2.6.12 kernel - what does colinux 0.8.x use? Can I expect 2.6.12 to play nicely with the debian 4 startup sequence, and udev on the fly device creation (I barely know what I am talking about). As I said i don't want to lose colinux's unique ability to let me run the very same linux installation i can also choose to boot natively. Ideally I'd free some disk space and install etch to a new parttion, but it's so damn hard to figure out the correct \Device\HarddiskX\PartitionY ! Also here there tentative release dates for the upcoming 0.7 and 0.8 versions? |
From: Henry N. <Hen...@Ar...> - 2007-04-17 14:26:51
|
Hello, baldyeti wrote: > Thanks Henri. But Woody and 2.4.x kernels are definitely > even older than the sarge I want to replace! Etch comes with > a 2.6.18 kernel and uses udev - that was my primary concern. > So I if read you correctly, colinux 0.7.0 will come with a > 2.6.12 kernel that's true. colinux 0.7.1 has kernel 2.6.12, udev is not enabled. > - what does colinux 0.8.x use? Can I expect > 2.6.12 to play nicely with the debian 4 startup sequence, > and udev on the fly device creation You should use coLinux 0.8.0, from snapshot 20070302 http://www.colinux.org/snapshots or http://www.henrynestler.com/colinux/testing/devel-0.8.0/20070302-Snapshot/ This is kernel 2.6.17 and udev is enabled. Newer kernel are not available. > Also here there tentative release dates for the upcoming 0.7 > and 0.8 versions? The 0.7.1 release hangs on a liddle bug while shutting down ('halt'). This is typicaly under Vista or some times under heavy memory usage in XP (XP with 256MB and coLinux configured mem=128 or more). Mostly if using initrd. The condition for the bug is very clear, I have no idea to fix. Grep in the devel list to read more. The 0.8.0 has very long time before release. You can still use the snapshots from http://www.colinux.org/snapshots, they tends to run good. -- Henry |
From: George P B. <geo...@gm...> - 2007-04-17 18:06:41
|
Henry Nestler wrote: > I'm not known about Debian versions and kernel depencies. > From my installations I know coLinux 0.6.4, 0.7.1 and 0.8.x runs on a > Debian Woody. > I recently tried to follow the debian install steps from the wiki to create a new etch image, but apparently there are additional requirements to get it to work for installing. I don't know at this point what those requirements are. I don't know if there's some command-line we can give it that will allow it to run without udev or if udev is going to be needed, or if it's an file system that we need support but don't have. I haven't gotten that far. I just know I wasn't able to follow the steps there to get it working. I was able to start with the iso and initrd set, but I wasn't able to mount them (I tried with aliases and without and I wasn't able to mount them, but haven't gotten so far as to creating the block devices necessary to see if that gets around that problem). This is for trying to great an etch image file, not for dual-booting, of course. > CoLinux versions 0.6.4, 0.7.1 and 0.8.0 runs also on my SuSE 9.0 dual > boot with an original 2.4.21 kernel. From my SuSE I know, that you can > run every coLinux on an older distry. For very old distries with kernel > 2.4.x you need to upgrade only the module-init-tools with fallback for > 2.4-kernels. > You saying that he SuSE original kernel was 2.4.21 or that you have an unofficial colinux patched 2.4.21 kernel? George |
From: Jun O. <ok...@di...> - 2007-04-17 18:15:35
|
Long time no see, folks. Sorry if I dont understand the matter correctly, but why you dont use debootstrap( or cdebootstrap) ? --- Okajima, Jun. Tokyo, Japan. |
From: David K. <da...@gi...> - 2007-04-17 19:24:45
|
Hi George, George P Boutwell <geo...@gm...> wrote: > I recently tried to follow the debian install steps from the wiki to > create a new etch image, but apparently there are additional > requirements to get it to work for installing. I don't know at this > point what those requirements are. I don't know if there's some > command-line we can give it that will allow it to run without udev or > if udev is going to be needed, or if it's an file system that we need > support but don't have. I haven't gotten that far. I just know I > wasn't able to follow the steps there to get it working. I was able > to start with the iso and initrd set, but I wasn't able to mount them > (I tried with aliases and without and I wasn't able to mount them, but > haven't gotten so far as to creating the block devices necessary to > see if that gets around that problem). This is for trying to great > an etch image file, not for dual-booting, of course. I'm running Debian etch on colinux -- last night I installed a fresh 0.7.1 colinux (with its 2.6.12-co-0.7.1 kernel) and the old debian image Debian-3.0r2.ext3-mit-backports.1gb.bz2 and then updated it to sarge (the old stable) and then to etch (the new stable). I changed /etc/apt/sources.list, commented out the backports lines, replaced "stable" with "sarge", ran "apt-get update" and "apt-get upgrade". This brought sarge up-to-date. Then I changed /etc/apt/sources.list again, replacing "sarge" with "etch", washed rinsed and repeated ("apt-get update" and "apt-get upgrade" again). I don't know how to make a snapshot out of this installation or otherwise share it with others, but it is *apparently* working just fne with the older kernel and no "udevs". I confess that I don't know what exactly a udev is though, or if mine is or isn't working :-) ...but the webserver, ssh and mysql all all working fine for me (I don't run X, access any of my windows USB devices from colinux or dual-boot though, if that matters). If someone could help me help others by creating a snapshot... I'd be glad to pitch in and share! -dave |
From: Wolfram W. <Wol...@ek...> - 2007-04-18 08:14:12
|
David Kaufman schrieb: > I changed /etc/apt/sources.list, commented out the backports lines,=20 > replaced "stable" with "sarge", ran "apt-get update" and "apt-get=20 > upgrade". This brought sarge up-to-date. Then I changed=20 > /etc/apt/sources.list again, replacing "sarge" with "etch", washed=20 > rinsed and repeated ("apt-get update" and "apt-get upgrade" again). so do i. but the result did not satisfy me. every time i want to install a new package (mc for example) i get the err= or=20 message from apt-get install <package>: E: This installation run will require temporarily removing the essential = package e2fsprogs due to a Conflicts/Pre-Depends loop. This is often bad,= =20 but if you really want to do it, activate the APT::Force-LoopBreak option= =2E E: Internal Error, Could not early remove e2fsprogs shoudn't be the e2fsprogs after the update/upgrade to etch the right vers= ion? --=20 Sch=F6ne Gr=FC=DFe aus Reutlingen Wolfram Wadepohl Forschung & Entwicklung E&K AUTOMATION Indumat GmbH & Co. KG Siemensstra=DFe 3 72766 Reutlingen Deutschland Tel. +49 7121 514-289 Fax +49 7121 514-299 eMail Wol...@ek... W.W...@in... W.W...@ie... WWW http://www.ek-automation.com http://www.indumat.de Diese Nachricht ist keine gesch=E4ftliche Mitteilung i. S. des EHUG. Bitte senden Sie mir keine Word- oder PowerPoint- (tm Microsoft) Anh=E4ng= e. Senden Sie mir einfachen Text, HTML oder PDF. Siehe http://www.gnu.org/philosophy/no-word-attachments.de.html |
From: baldyeti <e_...@ho...> - 2007-04-18 08:38:06
|
> I don't know how to make a snapshot out of this installation or > otherwise share it with others, but it is *apparently* working just fne > with the older kernel and no "udevs". I confess that I don't know what > exactly a udev is though, or if mine is or isn't working :-) Pretty mysterious to me too, hence my questions here ;-) I think I read that etch will only use udev for fresh installs, not when upgrading, which would explain what you're seeing. Ideally I'd do a fresh install to say hdc11. But then how do I find the proper device alias for colinux (the wiki instructions are kinda hazy). Other than that, i'll try to backup, then simply let the upgrade run and see what happens... |
From: George P B. <geo...@gm...> - 2007-04-18 23:03:27
|
baldyeti wrote: > Pretty mysterious to me too, hence my questions here ;-) > I think I read that etch will only use udev for fresh installs, > not when upgrading, which would explain what you're seeing. > Ideally I'd do a fresh install to say hdc11. But then how do > I find the proper device alias for colinux (the wiki instructions > are kinda hazy). Other than that, i'll try to backup, then simply > let the upgrade run and see what happens... > At the risk of getting way off-topic here... and with this 'disclaimer': this is MY UNDERSTANDING of it and it may not be complete or 100% accurate... udev is a replacement for devfs, which was weighed and found wanting. udev like devfs determines device nodes more or less dynamically by what the device drivers in the Linux kernel report or register back with the kernel. The major difference between devfs and udev is that udev is largely a user program with a much, much smaller kernel footprint (ie code that runs in the kernel/as a module). There where bugs in the early forms of udev, such that only more recent kernels (I believe magic kernel that most have settled on for 'proper' udev support is anything beyond 2.6.16). You need not only udev support in the kernel, but a udev program that runs early in the boot process. I believe that unless you particularly want udev, you can always bypass it/get around it by manually creating the device nodes that you need. Additionally, most the udev configurations allow 'static' or force devices, although those shouldn't be necessary for coLinux dives as they generally make the necessary calls to the kernel to 'register' with devfs and/or udev. How this all this relates to coLinux... newer distros that are requiring udev & forcing an reasonable kernel version will require coLinux 0.8.0-based kernels or the udev daemon (if they don't do a kernel version check) be configured to have the same device nodes as you'd have if you weren't using udev, or udev disabled. If they do kernel version checks, you will need to work around this (if possible) if not, then you have to use 0.8.0-based experimental version of your uses. Again this is my basic understanding of things and not a complete anthology, there is lots of information on the web on udev, configuring it, etc. For the real scoop, do a little light digging in this information and be enlightened. :) George |
From: David K. <da...@gi...> - 2007-04-18 15:56:25
|
Hi Wolfram, Wolfram Wadepohl <Wol...@ek...> wrote: >> I [...] replaced "stable" with "sarge", ran "apt-get update" and >> "apt-get upgrade" [...] washed rinsed and repeated.... > > so do i. but ... i get the error message: > E: This installation run will require temporarily removing the > essential package e2fsprogs due to a Conflicts/Pre-Depends loop. This > is often bad, but if you really want to do it, activate the > APT::Force-LoopBreak option. E: Internal Error, Could not early > remove e2fsprogs shoudn't be the e2fsprogs after the update/upgrade > to etch the right version? I got that too. The coLinux Wiki mentioned a workaround that worked around for me, too: http://colinux.wikia.com/wiki/DebianRootFsImages Also, the Debian etch Release Notes has a (very long but very informative) page on upgrade issues that helps with problems such as this: http://www.debian.org/releases/etch/i386/release-notes/ch-upgrading.en.html#s-trouble Hope this helps! -dave |
From: Henry N. <Hen...@Ar...> - 2007-04-17 20:04:32
|
George P Boutwell wrote: > Henry Nestler wrote: >> CoLinux versions 0.6.4, 0.7.1 and 0.8.0 runs also on my SuSE 9.0 dual >> boot with an original 2.4.21 kernel. From my SuSE I know, that you can >> run every coLinux on an older distry. For very old distries with kernel >> 2.4.x you need to upgrade only the module-init-tools with fallback for >> 2.4-kernels. >> > You saying that he SuSE original kernel was 2.4.21 or that you have an > unofficial colinux patched 2.4.21 kernel? no, no. ;-) I'm running only the coLinux 2.6.x kernels from 2.6.12 to 2.6.17 there. The module-init-tools have updated for my SuSE. The fallback is for native Linux boot with suSE kernel 2.4.21-default. -- Henry |