Re: Error installing Squeeze with root filesystem as btrfs
On 9 January 2011 09:20, Boyd Stephen Smith Jr. b...@iguanasuicide.net wrote: In aanlktimma2=mmo758p8xb85apjjqzv3+=mhukmoqt...@mail.gmail.com, Paul Richards wrote: On 8 January 2011 08:48, Boyd Stephen Smith Jr. b...@iguanasuicide.net wrote: In aanlktikebsmpx7jabaoeguptdp6bdkglyulphdb1p...@mail.gmail.com, Paul Richards wrote: On first boot into the system I got the following error: FATAL: Error inserting btrfs (/lib/modules/2.6.32-5-amd64/kernel/fs/btrfs/btrfs.ko): Unknown symbol in module, or unknown parameter (see dmesg) Make sure your initrd contains the crc32 module before you go looking for other problems. IIRC, this can be configured in /etc/iniramfs-tools/modules, but you may be using some other method to build an initrd. I'm not doing anything myself to build the initrd. I'm simply running the current ISO candidate for the squeeze netinst installer. Is what you suggest something I can fix from the terminal during the install? (I can alt-f2 to the spare VT to perhaps make the corrections) I did the migration to btrfs root from a working reiserfs root, so I'm not sure if the fix can be applied from the netinst iso. I'm not familiar enough with the netinst process to be sure, but I think you should be able to chroot, make the config changes, and rebuild the initrd just before the netinst reboots into the new system. I figured out how to add crc32 to the initramfs, but unfortunately I get the same error. It's unfortunate that the error message does not tell me which symbol is missing, as perhaps including another modulre in the initramfs would help. Since it took me a little time to figure out, I'll write out my procedure here. At the final prompt of the installer (where it asks you to remove the installation CD and prepare for reboot), I pressed alt-f2 to jump to the spare VT. From here I did the following: 1. Mount /dev /sys and /proc into /target/..: mount -o bind /dev /target/dev mount -o bind /sys /target/sys mount -o bind /proc /target/proc 2. Chroot into /target: chroot /target 3. Edit /etc/initramfs-tools/modules to contain crc32 (I also added btrfs for extra good measure) vi /etc/initramfs-tools/modules 4. Rebuild the initramfs dpkg-reconfigure linux-image-2.6.32-5-amd64 (you should see confirmation that grub was updated too) 5. Exit chroot, and alt-f1 back to the installer VT. -- Paul Richards @pauldoo -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktincxu0eczjy2ojdsojuyvzthzxnpxbhmpwu6...@mail.gmail.com
Re: Error installing Squeeze with root filesystem as btrfs
On 9 January 2011 12:37, Pascal Hambourg pascal.m...@plouf.fr.eu.org wrote: Paul Richards a écrit : I figured out how to add crc32 to the initramfs, but unfortunately I get the same error. It's unfortunate that the error message does not tell me which symbol is missing, as perhaps including another modulre in the initramfs would help. Did you look at the output of dmesg as the message suggested ? The initramfs should start a shell after failing to mount the root filesystem (be patient, this may take some time). In the installer/rescue shell, you could also check the module dependency with lsmod. I tried this procedure again but used a shotgun approach to picking which extra modules to put into the initramfs. From the installer VT (at step 3 in my procedure above) I executed: lsmod | sed 's%[ \t].*$%%' /etc/initramfs-tools/modules This has probably led to a bloated ramfs, but the system now boots with btrfs as the root filesystem. :) Thanks everyone for your help. -- Paul Richards @pauldoo -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlkti=dgoy=fmht5xzgmdkds1sjez21nwjqxfgjy...@mail.gmail.com
Re: Error installing Squeeze with root filesystem as btrfs
On 9 January 2011 13:45, Paul Richards paul.richa...@gmail.com wrote: On 9 January 2011 12:37, Pascal Hambourg pascal.m...@plouf.fr.eu.org wrote: Paul Richards a écrit : I figured out how to add crc32 to the initramfs, but unfortunately I get the same error. It's unfortunate that the error message does not tell me which symbol is missing, as perhaps including another modulre in the initramfs would help. Did you look at the output of dmesg as the message suggested ? The initramfs should start a shell after failing to mount the root filesystem (be patient, this may take some time). In the installer/rescue shell, you could also check the module dependency with lsmod. I tried this procedure again but used a shotgun approach to picking which extra modules to put into the initramfs. From the installer VT (at step 3 in my procedure above) I executed: lsmod | sed 's%[ \t].*$%%' /etc/initramfs-tools/modules This has probably led to a bloated ramfs, but the system now boots with btrfs as the root filesystem. :) Thanks everyone for your help. A simpler solution appears to be installing 'btrfs-tools' during install. This package drags in some initramfs-tools magic[1] that causes the correct modules to be put into the initramfs. This is far more elegant than hacking /etc/initramfs-tools/modules 1: http://packages.debian.org/squeeze/amd64/btrfs-tools/filelist -- Paul Richards @pauldoo -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktinedgbajk_z+ehs7ulvhqcgcignp55o6d9oc...@mail.gmail.com
Re: Error installing Squeeze with root filesystem as btrfs
On 8 January 2011 08:48, Boyd Stephen Smith Jr. b...@iguanasuicide.net wrote: In aanlktikebsmpx7jabaoeguptdp6bdkglyulphdb1p...@mail.gmail.com, Paul Richards wrote: On first boot into the system I got the following error: FATAL: Error inserting btrfs (/lib/modules/2.6.32-5-amd64/kernel/fs/btrfs/btrfs.ko): Unknown symbol in module, or unknown parameter (see dmesg) A few other errors follow as a consequence of not being able to mount the root filesystem. Is installing to a root btrfs filesystem expected to work in squeeze yet? I had it working under a mostly Lenny system. It continues to work although that laptop is mostly Squeeze now. There was an issue with initramfs-tools and btrfs on root. The btrfs module depends on libcrc32 module, which is correctly pulled in. However, some systems also need to crc32 module, if no other module provides the a crc32 implementation. libcrc32 is an interface to the crc32 module software or a hardware (assisted) implementation. Make sure your initrd contains the crc32 module before you go looking for other problems. IIRC, this can be configured in /etc/iniramfs-tools/modules, but you may be using some other method to build an initrd. I'm not doing anything myself to build the initrd. I'm simply running the current ISO candidate for the squeeze netinst installer. Is what you suggest something I can fix from the terminal during the install? (I can alt-f2 to the spare VT to perhaps make the corrections) -- Paul Richards @pauldoo -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktimma2=mmo758p8xb85apjjqzv3+=mhukmoqt...@mail.gmail.com
Error installing Squeeze with root filesystem as btrfs
I installed the Debian squeeze from the current beta 2 netdisk installer (debian-squeeze-di-beta2-amd64-netinst.iso). I configured an ext4 partition for /boot on /dev/sda1, and a btrfs partition for / on /dev/sda2. On first boot into the system I got the following error: FATAL: Error inserting btrfs (/lib/modules/2.6.32-5-amd64/kernel/fs/btrfs/btrfs.ko): Unknown symbol in module, or unknown parameter (see dmesg) A few other errors follow as a consequence of not being able to mount the root filesystem. Is installing to a root btrfs filesystem expected to work in squeeze yet? -- Paul Richards @pauldoo -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktikebsmpx7jabaoeguptdp6bdkglyulphdb1p...@mail.gmail.com
A warning how not to upgrade from Lenny to Squeeze
Hi, I recently upgraded from Lenny to Squeeze. I did the upgrade rather haphazardly without much preplanning and the procedure turned out to be a bit of a nightmare. I acknowledge that Squeeze is not fully released (just frozen), so I understand that there may not be any official upgrade guide yet (and I certainly imagine it would look different to what I attempted). So what follows is *not* a criticism of Debian, but I hope it might serve as a warning to others. :) (Perhaps it may help some technical guys debug the upgrade process, or perhaps it might help someone who is attempting a naive upgrade like I did.) Here's what I did.. I updated /etc/apt/sources.list = s/lenny/squeeze/ I then tried aptitude update aptitude install apt dpkg aptitude. This presented me with a solution that looked too broke for my taste, so I didn't apply it. I then tried aptitude safe-upgrade, this looked sensible so I gave it a shot. This fell over with a few package configuration errors, so I repeatedly tried aptitude safe-upgrade / aptitude full-upgrade in alternate shots. This iterated for a while with different errors each time until eventually I ended up with a system without aptitude installed. I then tried apt-get -f install, several times, which seemed to clear up a lot of things. Eventually my repeated apt-get -f install commands were just giving me the same udev error message. Apparently udev and the kernel need to be updated at the same time, but none of the tools would let me install the new kernel and udev (too many broken packages at the same time I think). I found a magic file that I could touch (/etc/udev/kernel-upgrade) to cause the udev config step to go ahead regardless (even though it knew my kernel wasn't up to it). I then did more rounds of apt-get -f install to clear up some more pending mess. I now had udev but no corresponding kernel to work with it. Now would have been a very bad time to reboot. :) I then manually used apt-get install to ensure that I had the new udev compatible kernel (linux-image-2.6-amd64 in my case).. I then went back to aptitude full-upgrade for a few more iterations, and eventually the system converged with no errors. After a reboot the system was pretty much fine except that X freaked out. I had to just empty the xorg.conf file, and it seems happy again now. Everything else like the new boot scripts, upgrade to grub2, new kernel, etc are working great. -- Paul Richards @pauldoo -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktinnylyberhshkjmq2bebq5szgud5huadwe5o...@mail.gmail.com
Forming drives together into a cache hierarchy
Is there a way to combine a large slow drive with a fast small drive in such a way that the faster drive simply becomes a cache for the larger drive? I imagine a computer where I'd like to have a small fast SSD and a large but slow spinning disk. I'd like to put the two block devices together so that I'm presented with a single logical block device, where the operating system is using the SSD as a cache (either inclusive or exclusive, write-though or otherwise) of the spinning disk. Is there some magic in LVM that can do this? If there is nothing at the block device level can it be done at the filesystem level? -- Paul Richards -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Forming drives together into a cache hierarchy
2009/9/2 Paul Richards paul.richa...@gmail.com: Is there a way to combine a large slow drive with a fast small drive in such a way that the faster drive simply becomes a cache for the larger drive? I imagine a computer where I'd like to have a small fast SSD and a large but slow spinning disk. I'd like to put the two block devices together so that I'm presented with a single logical block device, where the operating system is using the SSD as a cache (either inclusive or exclusive, write-though or otherwise) of the spinning disk. Is there some magic in LVM that can do this? If there is nothing at the block device level can it be done at the filesystem level? This looks close: http://www.cis.fiu.edu/~zhaom/dmcache But unfortunately this isn't available in a standard Debian installation. Perhaps I simply need to wait a while longer. -- Paul Richards -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Forming drives together into a cache hierarchy
2009/9/2 Jochen Schulz m...@well-adjusted.de: Ron Johnson: On 2009-09-02 05:30, Paul Richards wrote: Is there a way to combine a large slow drive with a fast small drive in such a way that the faster drive simply becomes a cache for the larger drive? -- snip I'd think about selling the SSD and either buy more RAM or a SCSI controller and a 10K RPM drive. While I would never recommend to sell an SSD (only to buy a larger one :)), I think buying more RAM is probably the way to go. Nothing beats the filesystem cache, not even an SSD. My recommendation: buy more RAM, install the OS to an SSD and copy the data most often used there as well. Put the rest on a regular disk and write a boot script to force the data you are interested in into the filesystem cache. If you cannot tell which data you are interested in in advance or if this subset of your data changes often, caching on an SSD is a hard problem to solve anyway. I have to admit I'm slightly surprised by your response. Caching has massive benefit for main memory, so much so that we have several layers of cache. What is different about hard disks? I don't see why RAM (filesystem cache) - SSD - Rotating disk is such a bad idea.. An SSD cache has a number of benefits over a filesystem cache in RAM. Price per GB will be much lower, and also it will persist over a reboot. -- Paul Richards -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Virtualbox Lenny Guest - Screen resize and Seamless mode
2009/8/3 Chris list.hursch...@gmx.de: On Monday 03 August 2009, go...@dobosevic.com wrote: Chris wrote: I am trying to get Screen resize and Seamless mode working on Lenny guest systems to no avail. I have tried insalling the guest-utils packages, as well as the guest-additions the manual way with m-a. They install fine, but sreen-resize and seamless mode don't work. Does anyone have this working at all? I've tried various xorg.conf files with the video driver and no screen resolutions. I am running virtualbox-ose 2.1.4 from lenny-backports on the host. Thanks, Chris I also have same problem. I have Debian Sid LXDE, Debian Sid XFCE and Debian Sid KDE as a guest. They were first Lenny and screen size was to big so I was edit xorg.conf and it was fine. After upgrade to Sid all of them have to small screen with black frame around (800x600 on 15 laptop, native resolution is 1024x768). Editing xorg.conf doesn't help any more. Bye, Goran Dobosevic Hrvatski: www.dobosevic.com English: www.dobosevic.com/en/ Lets see if anyone has it working! I have some SIDUX guests (09-01, 09-02) that work... It works for me with Virtual Box 3.0.2 and Debian squeeze. It took me a little while to get it working properly though, I can take a look later when I get access to the machine to see what my final configuration was. I remember experimenting with different versions of the guest additions. I tried installing the debian packages, I tried installing the guest additions from the Virtual Box menu, and I tried just nothing (xorg already has some built-in support). Unfortunately I can't remember which I settled on. -- Paul Richards -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: How to partition and format external disk to ext3
2009/8/3 Paul E Condon pecon...@mesanetworks.net: To discover the device name and partition name that the kernel wants you to use for the HDD, run 'ls -l /dev/sd*' If this give more than two lines of output, unplug whatever other USB devices you are running and repeat. If you still have more than two lines, go slow, and think hard. Try unplugging your new HDD and plugging it in again, and monitor how/what device names change as you do that. An alternative is to run 'dmesg' shortly after plugging in the new drive. It'll spit out a series of messages, the tail end of which should show the new drive being recognised. Another alternative is to run 'mount' priort to ejecting/unmounting the disk from your desktop environment. You'll see from the listing the name of the USB drive. -- Paul Richards -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: new wireless card
I can recommend that you don't get a Ralink RT2561. I have one, and while it works fine it needs binary drivers[1]. :( Unfortunately I didn't realise this at the time I bought. 1: http://wiki.debian.org/rt61pci 2009/4/10 steef debian.li...@home.nl: hello to you all out there, my asus_wireless ethernetcard, chip rt2500 (ralink), is of no use any more. please can somebody tell me which is the best lenny_compatible wireles ethernet card nowadays? thank you very much, regards, steef -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org -- Paul Richards -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: new wireless card
2009/4/10 Kelly Clowers kelly.clow...@gmail.com: On Thu, Apr 9, 2009 at 23:14, steef debian.li...@home.nl wrote: please can somebody tell me which is the best lenny_compatible wireles ethernet card nowadays? On Fri, Apr 10, 2009 at 01:14, Paul Richards paul.richa...@gmail.com wrote: I can recommend that you don't get a Ralink RT2561. I have one, and while it works fine it needs binary drivers[1]. :( Unfortunately I didn't realise this at the time I bought. 1: http://wiki.debian.org/rt61pci I don't see anything about binary drivers there, just the usual firmware issues. Lots of devices need loadable firmware that is not Free. It is a very hard issue that the community is really just beginning to tackle - some don't even think it is an issue that needs to be worried about! Yes sorry, my mistake. It's the firmware that is not Free. -- Paul Richards -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Upgrading from linux-image-2.6.18-686 to linux-image-2.6.24-686 breaks on AMD Geode LX CPU
2009/4/9 Sven Hoexter s...@timegate.de: On Thu, Apr 09, 2009 at 04:03:18PM +0100, Paul Richards wrote: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=464962 Turns out the AMD Geode LX is affected by the same problem, and there is already a fix in the pipeline. I've just taken a fresh look at this to see if I'm now able to move back to a normal 686 kernel, but I see that the fixed version 2.6.26-8 is not available on Lenny. Currently Lenny still has 2.6.26-1. Is there any chance this will become available on Lenny eventually? Hm currently there's -13 with -13lenny2 scheduled for the next stable release available for Lenny. s...@arthur:~$ rmadison linux-image-2.6.26-1-686 linux-image-2.6.26-1-686 | 2.6.26-13 | stable | i386 linux-image-2.6.26-1-686 | 2.6.26-13 | testing | i386 linux-image-2.6.26-1-686 | 2.6.26-13lenny2 | proposed-updates | i386 If you've still something older than that you should check your system and maybe the mirror why this is the case. I don't have proposed-updates in my sources.list. Is it advisable to add proposed-updates for a stable system? Personally I decided to run my Geode LX system with the 486 kernel image. Donno how much of a performance penalty there is but I guess for usual system usage it doesn't matter. Likewise, I can't imagine that it makes a significant difference. It would just make my Geode system sound more powerful if the login message splurted 686 instead of 486.. :) -- Paul Richards -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Upgrading from linux-image-2.6.18-686 to linux-image-2.6.24-686 breaks on AMD Geode LX CPU
2008/4/9 Paul Richards paul.richa...@gmail.com: 2008/4/8 Paul Richards paul.richa...@gmail.com: Hi, I recently upgraded from etch to lenny and in the process my kernel was upgraded from linux-image-2.6.18-686 to linux-image-2.6.24-686. This new kernel crashes immediately after being uncompressed on boot, and I have been forced to use the 486 build of the kernel if I want 2.6.24. This AMD Geode LX processor has always executed 686 kernels in the past. Have I previously just been lucky or is this a bug in 2.6.24? A friend of mine pointed me towards this: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=464962 Turns out the AMD Geode LX is affected by the same problem, and there is already a fix in the pipeline. I've just taken a fresh look at this to see if I'm now able to move back to a normal 686 kernel, but I see that the fixed version 2.6.26-8 is not available on Lenny. Currently Lenny still has 2.6.26-1. Is there any chance this will become available on Lenny eventually? -- Paul Richards -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Upgrading from linux-image-2.6.18-686 to linux-image-2.6.24-686 breaks on AMD Geode LX CPU
2008/4/8 Paul Richards [EMAIL PROTECTED]: Hi, I recently upgraded from etch to lenny and in the process my kernel was upgraded from linux-image-2.6.18-686 to linux-image-2.6.24-686. This new kernel crashes immediately after being uncompressed on boot, and I have been forced to use the 486 build of the kernel if I want 2.6.24. This AMD Geode LX processor has always executed 686 kernels in the past. Have I previously just been lucky or is this a bug in 2.6.24? A friend of mine pointed me towards this: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=464962 Turns out the AMD Geode LX is affected by the same problem, and there is already a fix in the pipeline. -- Paul Richards -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Upgrading from linux-image-2.6.18-686 to linux-image-2.6.24-686 breaks on AMD Geode LX CPU
Hi, I recently upgraded from etch to lenny and in the process my kernel was upgraded from linux-image-2.6.18-686 to linux-image-2.6.24-686. This new kernel crashes immediately after being uncompressed on boot, and I have been forced to use the 486 build of the kernel if I want 2.6.24. This AMD Geode LX processor has always executed 686 kernels in the past. Have I previously just been lucky or is this a bug in 2.6.24? This was the only problem I encountered during the etch to lenny upgrade, so congratulations to everyone involved! Below is my /proc/cpuinfo. [EMAIL PROTECTED]:~$ cat /proc/cpuinfo processor : 0 vendor_id : AuthenticAMD cpu family : 5 model : 10 model name : Geode(TM) Integrated Processor by AMD PCS stepping: 2 cpu MHz : 498.066 cache size : 128 KB fdiv_bug: no hlt_bug : no f00f_bug: no coma_bug: no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu de pse tsc msr cx8 sep pge cmov clflush mmx mmxext 3dnowext 3dnow bogomips: 997.66 clflush size: 32 -- Paul Richards -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]