Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
FS#15281 - [kernel26] Boot fails after upgrade to kernel 2.6.30-5
Attached to Project:
Arch Linux
Opened by Ray Clancy (lilsirecho) - Saturday, 27 June 2009, 02:39 GMT
Last edited by Roman Kyrylych (Romashka) - Saturday, 25 July 2009, 16:18 GMT
Opened by Ray Clancy (lilsirecho) - Saturday, 27 June 2009, 02:39 GMT
Last edited by Roman Kyrylych (Romashka) - Saturday, 25 July 2009, 16:18 GMT
|
DetailsDescription:Upgrade to kernel "30-5-x86_64" on two usb HDD, one ext3 the other ext4 produces same failure to boot stating "unable to detect root device uuid(not reprinted herein)
Request to raise ..."rootdelay=10 or greater.This is done with edit but result is the same. Fallback produces same result. Occurs on both usb HDD's after upgrade. Additional info: * package version(s) * config and/or log files etc. Steps to reproduce:I have no idea as to what steps will reproduce this failure to boot. My arch HDD's cannot boot. I have no experience with chroot in x86_64 systems and would not appreciate trying any i686 procedures for fear of making things worse. A procedure to chroot into my hdd's which enables a return to the previous kernel might be possible. However, many klibc , module-init-tools,mkinitcpio and kernel26-firmware were installed in the upgrade along with the kernel. I cannot guess which caused me to lose boot capability. It may have been a bad install? A proper x86_64 chroot procedure is a good start ...especially if the packages needed to be downgraded are outlined. |
This task depends upon
Tried booting the new install and received the same failure as reported for the two other drives.
One difference...the usb devices remained enabled to allow ..."reboot" from "ranfs$ recovery which did not occur with the subject hdd's...usb failed on those drives since the keyboard and mouse devices were disabled(lights went out).
Thus, it seems a problem exists with the kernel and its companion mkinitcpio and klibc packages.
Entering ...rootdelay=8,10,16 each in turn gave no change.
The root uuid is not detected. Maybe its udev or autodetect?
http://bbs.archlinux.org/viewtopic.php?pid=575922#p575922
Some initrd activity is shutting down usb HDD so the root device is not recognized.
FS#15282@Ray: see the comment in that report.
Additional comment on my bug report. The install of kernel "30" in the new drive with x86_64 was successful when using an external usb device but it failed to boot when using that device.
With the HDD installed in another external device, the system booted. This indicates that the external devices are causing the problem by not responding to boot activity.
I note that the successful boot sequence shows USB shuts down, restarts,shuts down and then restarts to remain enabled for the completion of the boot. This occurs during initrd.
The drives which do not boot do not display that usb activity. The usb during initrd shuts off and stays off. Since my system uses usb boot, it fails to boot.
No help occurs with rootdelay changes.
find: /sys/devices....no such file or directory.........
cd /etc............nano mkinitcpio.conf...removed hook "autodetect" saved...
ran mkinitcpio -p kernel26
exit
reboot
Usb still stays disabled during initrd.
Tried fallback...same result..no usb
However, an upgraded Western Digital HDD to kernel "30" cannot boot in the same enclosure because USB is disabled during initrd.
What is the difference between an upgrade and a new install to the same kernel?
Is it caused by the new usb-start-ehci-first.conf which was not actually active in kernel "29"?
Why should all usb be disabled and stay disabled? Most of initrd doesn't run due to loss of usb (almost immediately in the sequence).
Is this an enclosure problem or a HDD problem or a kernel problem?
Tried many times to correct things...mounted root,boot,/proc/sys/tmp all into /mnt...chroot /mnt......ran mkinitcpio -p kernel26.....exit reboot.....still fail due to no usb.
Why no usb?
Arch is not "simple" to repair after upgrade oops..................
Both original HDD and backup do not usb boot after upgrade to kernel "30".
I suspect not many users use usb boot............
The "ehci first" modprobe file is not new, it has been there for a while, it just needed to be renamed due to a change in modprobe. This file fixes some USB bugs, but shouldn't cause any.
You're saying that you tried both hard drives from the same USB enclosure, so there must be a difference in the initramfs images - I don't see what this difference might be though.
I am not sure what you mean by emergency shell...is it ..ramfs$ ...? If so, that is dead as a doornail with no usb for keyboard and mouse.........
Also are there any messages displayed on the console when udev is started (and modules should be loaded)?
If any messages are displayed I cannot access them because all activity is disabled.
Can you compare the file list of both initramfs images (zcat /boot/kernel26.img | cpio -t). Both should contain at least ehci_hcd, one of ohci_hcd or uhci_hcd, usb-storage, scsi_mod and sd
Am I to do this compare using chroot into failed hdd? I tried lsusb and it didn't do diddly in chroot.
Right, no usb means on keyboard, this limits our recovery capability a bit.
> Messages on the console
If you omit the "quiet" option from the kernel commandline (it's not there by default), you should see all messages about hardware and potential hardware problems (USB timeouts and such). If everything is quiet after "Loading udev ..." then something is wrong. It's possible everything scrolls by too fast though-
> Comparing initramfs
You can just use the above command on any initramfs image (it's just a gzipped cpio archive), it doesn't matter which system is actually booted. Just boot any system that works and compare the filelists of the initramfs image on the working and non-working system.
I assume you mean I can boot with the new install hdd but have the non-booting hdd connected in usb enclosure and run the zcat...but how to run it on the non-booting hdd?
I will try the "quiet" option but expect no new revelation.
zcat on this booting HDD gave a long listing which included the ehci, uhci, usb-storage modules. I don't know if it will work on the non-booting hdd without a chroot.
This will take some time......
Supposedly kernel "30" eliminates the rootdelay= parameter.
When in kernel "29" I utilized rootdelay=8 to permit booting and entered that in /etc/grub.
I cannot understand why the error message requests entering rootdelay=10 or greater if mkinitcpio for kernel "30" eliminates rootdelay= as an option.
The kernel command line entry in boot has rootdelay=8 since it was entered in kernel "29" into grub.
Can this entry be the cause of failure to boot?
If so, why is there no warning to remove same from the grub file?
This would explain why the new installation boots and the upgrade doesn't.
Please access the initramfs from the non-working hard drive and check for the presence of the required modules there. The booting one seems to be alright with ehci, uhci and usb-storage (Caution: Some motherboards use uhci, some use ohci. If you have autodetection enabled in mkinitcpio, the hard drive will only boot on a machine with the same controller that the initramfs was built with. If you disable autodetection, both ohci and uhci will be present and it will boot on all machines).
As to finding the reason for your problem, I still have nothing, I'll keep thinking.
One item of interest is that the kernel-headers are "29" type in the non-booting hdd.
I assume removing autodetect from mkinicpio.conf is what you intend?
That's no problem, although with the new mkinitcpio, it is not necessary anymore.
> One item of interest is that the kernel-headers are "29" type in the non-booting hdd.
The "kernel-headers" package, despite its name, has nothing to do with the kernel. Those are actually the headers that glibc has been built against, and they must not be changed unless a new glibc is built against them.
> I assume removing autodetect from mkinicpio.conf is what you intend?
You can leave it in there. The fallback image is identical to the normal image, except that fallback is built without the autodetect hook. If something doesn't work with the normal image, you always still have the fallback image without autodetection.
To satisfy my curiosity, can you upload both kernel26.img files (from the working and the broken hdd) to the bug tracker? Also paste the kernel-line from each grub configuration. This way we can finally find out if this is a kernel bug or a bug in the initramfs.
/dev
/dev/null
/dev/zero
/dev/console
/dev/mem
/lib
/lib/klibc-UqSadMgryalzKq_XarP9XnQvbXQ.so
/bin
/bin/cat
/bin/chroot
/bin/dd
/bin/dmesg
/bin/false
/bin/fstype
/bin/halt
/bin/ipconfig
/bin/kill
/bin/kinit
/bin/ln
/bin/lodel
/bin/losetup
/bin/lsmod
/bin/mdassemble
/bin/minips
/bin/mkdir
/bin/mkfifo
/bin/mknod
/bin/moddeps
/bin/mount
/bin/mv
/bin/nfsmount
/bin/nuke
/bin/parseblock
/bin/pivot_root
/bin/poweroff
/bin/readlink
/bin/reboot
/bin/replace
/bin/resolve-modalias
/bin/resume
/bin/run-init
/bin/sh
/bin/sleep
/bin/sync
/bin/true
/bin/umount
/bin/uname
/sbin
/sbin/depmod
/sbin/insmod
/sbin/modinfo
/sbin/modprobe
/sbin/rmmod
:5492 blocks
/bin/modprobe
/init_functions
/init
/config
/etc
/etc/modprobe.d
/etc/modprobe.d/usb-load-ehci-first.conf
/sbin/udevd
/sbin/udevadm
/lib/udev
/lib/udev/rules.d
/lib/udev/rules.d/50-udev-default.rules
/lib/udev/rules.d/60-persistent-storage.rules
/lib/udev/rules.d/64-device-mapper.rules
/lib/udev/rules.d/64-md-raid.rules
/lib/udev/rules.d/80-drivers.rules
/lib/udev/firmware.sh
/lib/udev/path_id
/lib/udev/vol_id
/lib/udev/load-modules.sh
/etc/udev
/etc/udev/udev.conf
/hooks
/hooks/udev
/lib/modules
/lib/modules/2.6.30-ARCH
/lib/modules/2.6.30-ARCH/kernel
/lib/modules/2.6.30-ARCH/kernel/drivers
/lib/modules/2.6.30-ARCH/kernel/drivers/scsi
/lib/modules/2.6.30-ARCH/kernel/drivers/scsi/scsi_mod.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/ata
/lib/modules/2.6.30-ARCH/kernel/drivers/ata/libata.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/ata/pata_acpi.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/ata/ata_generic.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/ata/ata_piix.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/scsi/sd_mod.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/block
/lib/modules/2.6.30-ARCH/kernel/drivers/block/floppy.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/usb
/lib/modules/2.6.30-ARCH/kernel/drivers/usb/core
/lib/modules/2.6.30-ARCH/kernel/drivers/usb/core/usbcore.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/usb/host
/lib/modules/2.6.30-ARCH/kernel/drivers/usb/host/ehci-hcd.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/usb/host/uhci-hcd.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/usb/storage
/lib/modules/2.6.30-ARCH/kernel/drivers/usb/storage/usb-storage.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/cdrom
/lib/modules/2.6.30-ARCH/kernel/drivers/cdrom/cdrom.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/scsi/sr_mod.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-a4tech.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-apple.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-belkin.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-cherry.ko
:
/lib/modules/2.6.30-ARCH/kernel/drivers/scsi/sd_mod.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/block
/lib/modules/2.6.30-ARCH/kernel/drivers/block/floppy.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/usb
/lib/modules/2.6.30-ARCH/kernel/drivers/usb/core
/lib/modules/2.6.30-ARCH/kernel/drivers/usb/core/usbcore.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/usb/host
/lib/modules/2.6.30-ARCH/kernel/drivers/usb/host/ehci-hcd.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/usb/host/uhci-hcd.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/usb/storage
/lib/modules/2.6.30-ARCH/kernel/drivers/usb/storage/usb-storage.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/cdrom
/lib/modules/2.6.30-ARCH/kernel/drivers/cdrom/cdrom.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/scsi/sr_mod.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-a4tech.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-apple.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-belkin.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-cherry.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-chicony.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-cypress.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/input
/lib/modules/2.6.30-ARCH/kernel/drivers/input/ff-memless.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/usbhid
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/usbhid/usbhid.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-drff.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-ezkey.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-gaff.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-gyration.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-kensington.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-kye.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-logitech.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-microsoft.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-monterey.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-ntrig.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-petalynx.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-pl.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-samsung.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-sony.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-sunplus.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-tmff.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-topseed.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/hid/hid-zpff.ko
/lib/modules/2.6.30-ARCH/kernel/fs
/lib/modules/2.6.30-ARCH/kernel/fs/mbcache.ko
/lib/modules/2.6.30-ARCH/kernel/fs/ext2
/lib/modules/2.6.30-ARCH/kernel/fs/ext2/ext2.ko
/lib/modules/2.6.30-ARCH/kernel/fs/jbd
/lib/modules/2.6.30-ARCH/kernel/fs/jbd/jbd.ko
/lib/modules/2.6.30-ARCH/kernel/fs/ext3
/lib/modules/2.6.30-ARCH/kernel/fs/ext3/ext3.ko
/lib/modules/2.6.30-ARCH/modules.dep
/lib/modules/2.6.30-ARCH/modules.alias
/lib/modules/2.6.30-ARCH/modules.symbols
(END)
This is the booting HDD...Had to do it in segments..............
Not sure how to get the non-booting data since I have to chroot and may not get to copy it outa there. Only path there is usb, possibly.
I will have to type in data in a new comment for the kernel lines.
kernel /vmlinuz26 root=/dev/disk/by-uuid/6608f512-df06-461b-aeaa-8b35f8ad47b5 ro
kernel /vmlinuz26 root=/dev/disk/by-uuid/7eb6d9ce-5ae6-4696-a1be-de58597f7235 ro rootdelay=8
Cannot obtain the zcat data in any way I know about on the non-booting drive. The data looks very good for ehci ohci and uhci.
Perhaps I misunderstand what is to be done?
Is there another method of copying from a non-booted hdd? Usually root is locked when opening a usb hdd that is mounted under another boot device.
I guess you mean I should copy /boot/kernel26.img ?
/sys
/dev
/dev/null
/dev/zero
/dev/console
/dev/mem
/lib
/lib/klibc-UqSadMgryalzKq_XarP9XnQvbXQ.so
/bin
/bin/cat
/bin/chroot
/bin/dd
/bin/dmesg
/bin/false
/bin/fstype
/bin/halt
/bin/ipconfig
/bin/kill
/bin/kinit
/bin/ln
/bin/lodel
/bin/losetup
/bin/lsmod
/bin/mdassemble
/bin/minips
/bin/mkdir
/bin/mkfifo
/bin/mknod
/bin/moddeps
/bin/mount
/bin/mount
/bin/mv
/bin/nfsmount
/bin/nuke
/bin/parseblock
/bin/pivot_root
/bin/poweroff
/bin/readlink
/bin/reboot
/bin/replace
/bin/resolve-modalias
/bin/resume
/bin/run-init
/bin/sh
/bin/sleep
/bin/sync
/bin/true
/bin/umount
/bin/uname
:/bin/uname
/sbin
/sbin/depmod
/sbin/insmod
/sbin/modinfo
/sbin/modprobe
/sbin/rmmod
/bin/modprobe
/init_functions
/init
/config
/etc
/etc/modprobe.d
/etc/modprobe.d/usb-load-ehci-first.conf
/sbin/udevd
/sbin/udevadm
/lib/udev
/lib/udev/rules.d
/lib/udev/rules.d/50-udev-default.rules
:
/lib/udev/rules.d/50-udev-default.rules
/lib/udev/rules.d/60-persistent-storage.rules
/lib/udev/rules.d/64-device-mapper.rules
/lib/udev/rules.d/64-md-raid.rules
/lib/udev/rules.d/80-drivers.rules
/lib/udev/firmware.sh
/lib/udev/path_id
/lib/udev/vol_id
/lib/udev/load-modules.sh
/etc/udev
/etc/udev/udev.conf
/hooks
/hooks/udev
/lib/modules
/lib/modules/2.6.30-ARCH
/lib/modules/2.6.30-ARCH/kernel
/lib/modules/2.6.30-ARCH/kernel/drivers
/lib/modules/2.6.30-ARCH/kernel/drivers/scsi
/lib/modules/2.6.30-ARCH/kernel/drivers/scsi/scsi_mod.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/scsi/scsi_mod.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/ata
/lib/modules/2.6.30-ARCH/kernel/drivers/ata/libata.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/ata/pata_acpi.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/ata/ata_generic.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/ata/ata_piix.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/scsi/sd_mod.ko
/lib/modules/2.6.30-ARCH/kernel/drivers/block
/lib/modules/2.6.30-ARCH/kernel/drivers/block/floppy.ko
/lib/modules/2.6.30-ARCH/kernel/fs
/lib/modules/2.6.30-ARCH/kernel/fs/mbcache.ko
/lib/modules/2.6.30-ARCH/kernel/fs/ext2
/lib/modules/2.6.30-ARCH/kernel/fs/ext2/ext2.ko
/lib/modules/2.6.30-ARCH/kernel/fs/jbd
/lib/modules/2.6.30-ARCH/kernel/fs/jbd/jbd.ko
/lib/modules/2.6.30-ARCH/kernel/fs/ext3
/lib/modules/2.6.30-ARCH/kernel/fs/ext3/ext3.ko
/lib/modules/2.6.30-ARCH/modules.dep
/lib/modules/2.6.30-ARCH/modules.alias
:
lib/modules/2.6.30-ARCH/modules.symbols
:
HOOKS="base udev autodetect pata scsi sata filesystems"
I swear they were there originally.
I have tried so many things that I can't even recall.
I will see what is in mkinitcpio.conf niw......
HOOKS='base udev autodetect pata scsi usb usbinput filesystems'
The result was success with no comments or errors. Fallback also success.
The /boot/kernel.img is still missing about 200mb (compared to the booting hdd).
Where goeth the missing items?
The problem initially was probably caused by a defective usb hdd enclosure, this enclosure does not permit booting with this hdd but another enclosure permits booting.
I appreciate the support given by Braino in detecting the cause of the loss of usb.
I consider the problem solved as a hardware-induced problem.
this can be closed