[fedora-arm] Fwd: Help!Failed to find a suitable stage1 device when using PXE installation.
David A. Marlin
dmarlin at redhat.com
Tue Dec 30 16:14:47 UTC 2014
On 12/29/2014 08:10 PM, Brooks Hu wrote:
> I suppose the kickstart file was loaded successfully, because after
> changing content in kickstart file(for example, text mode), the
> installation process will reflect the change. You mentioned checking
> server logs, how can i do this? I have only access to a serial
> console. Thanks a lot.
When I refer to the 'server logs' I am referring to the logs on the
DHCP/TFTP/HTTP server(s) where the PXE configuration, install tree, and
kickstart configuration files are stored.
You can also inspect the anaconda logs on the target (Mustang) system
from the serial console. Anaconda uses tmux for the serial terminal
[anaconda] 1:main* 2:shell 3:log 4:storage-log 5:program-log
so you can toggle between the main screen (Ctrl-B 1) and a shell (Ctrl-B
2), or to the log files themselves (Ctrl-B and 3, 4, or 5), much as you
would switch screens (tabs) using 'screen' on a normal terminal window.
"Ctrl-B [" will put the active screen in 'copy mode' so you can scroll
up and down using the arrow keys to see more of the logs. The <Esc>
will exit copy mode.
> On Mon, Dec 29, 2014 at 11:56 PM, David A. Marlin <dmarlin at redhat.com
> <mailto:dmarlin at redhat.com>> wrote:
> On 12/29/2014 12:05 AM, Brooks Hu wrote:
>> Thanks for the help.
>> Regarding my configuration file, i just copied it from other thus
>> some lines wrong, but i think they have nothing to do with this
>> issue. This issue should only be related to disk partitioning.
> Are you certain that anaconda is finding the kickstart file
> (checked the server logs)? There were two issues pointed out before:
> 4) [!] Software selection (Nothing selected)
> 5) [!] Installation Destination (No disks selected)
> Since it did not know which Software was selected (%packages in
> the kickstart) or the Installation Destination (specified as sda
> in the kickstart), it may not be finding the kickstart at all, or
> may not be able to successfully parse it. Is it also possible
> that the initrd does not include all the requisite drivers to find
> the network or disk. Please inspect the anaconda logs to see what
> errors are occurring. From the shell window you can also verify
> that the network is functional (use curl to get the kickstart
> file), and check that the expected partitions are available.
>> I tried your kickstart configuration, but the issue still exists.
>> I noticed in the configuration there is a line "bootloader
>> --location=mbr", does it mean you are using legacy BIOS rather
>> than EFI? I am using EFI BIOS.
> I am using EFI as well. Please be sure you are using the most
> recent firmware version.
>> I downloaded the image from here
>> then setup a local FTP to host it.
> RC7 should be a good release to install, so I suspect is it an
> issue in the setup as mentioned above.
>> Best Regards,
>> On Mon, Dec 29, 2014 at 9:19 AM, David A. Marlin
>> <dmarlin at redhat.com <mailto:dmarlin at redhat.com>> wrote:
>> On 12/27/2014 06:19 PM, Brooks Hu wrote:
>>> ---------- Forwarded message ----------
>>> From: *Brooks Hu* <brooks.hu at gmail.com
>>> <mailto:brooks.hu at gmail.com>>
>>> Date: Friday, December 26, 2014
>>> Subject: Help!Failed to find a suitable stage1 device when
>>> using PXE installation.
>>> To: arm at lists.fedoraproject.org
>>> <mailto:arm at lists.fedoraproject.org>
>>> Hi all,
>>> I am trying automatic PXE installation (Fedora 21 for
>>> AARCH64) on my Mustang board, but looks like kickstart can't
>>> Here comes the detail:
>>> Starting installer, one moment...
>>> find_file: stat /proc/device-tree/chosen/bootpath, No such
>>> file or directory
>>> anaconda 21.48.21-1 for Fedora-Server 21 started.
>>> * installation log files are stored in /tmp during the
>>> * shell is available on TTY2
>>> * when reporting a bug add logs from /tmp as separate
>>> text/plain attachments
>>> 10:26:29 Not asking for VNC because of an automated install
>>> Starting automated install......
>>> Generating updated storage configuration
>>> storage configuration failed: failed to find a suitable
>>> stage1 device
>>> 1) [x] Language settings 2) [!] Timezone settings
>>> (English (United States)) (Timezone is not set.)
>>> 3) [x] Installation source 4) [!] Software
>>> (ftp://192.168.1.1/F21/) (Nothing
>>> 5) [!] Installation Destination 6) [x] Network
>>> (No disks selected) (Wired (eth0)
>>> 7) [!] Root password 8) [!] User creation
>>> (Password is not set.) (No user
>>> will be created)
>>> Not enough space in filesystems for the current software
>>> selection. An additional 2861.02 MiB is needed.
>>> Please make your choice from above ['q' to quit | 'b' to
>>> begin installation |
>>> UEFI version is 1.1.0-rh-0.13
>>> Here comes the kick start config file(I modified based on a
>>> sample from certain website):
>> Note that the installation options indicate: 4) [!] Software
>> selection (Nothing selected) and 5) [!] Installation
>> Destination (No disks selected), which implies anaconda is
>> not seeing (or accepting) some of the options in your
>> kickstart. Please look in the logs in /tmp to find any errors.
>> There are some things in you kickstart that I am not familiar
>> with, such as:
>> vznetcfg --net=virt_network1:eth0
>> vztturlmap $FS_SERVER http://myrepository.com
>> %eztmplates --cache
>> and the following just look wrong for an aarch64 install:
>> I would try a more "standard" Fedora kickstart install first
>> to make sure the repos, etc. are working, and then try adding
>> the 'custom' config options to see what breaks.
>> Attached is an example kickstart I have successfully used for
>> provisioning a Mustang board.
>>> lang en_US.UTF-8
>>> keyboard us
>>> part efi --fstype=efi --size=300 --ondisk=sda
>>> part /boot --fstype=ext4 --size=512 --ondisk=sda
>>> part / --fstype=ext4 --size=20096 --ondisk=sda
>>> part /vz --fstype=ext4 --size=40768 --ondisk=sda
>>> part swap --size=4000
>>> bootloader --location=partition --ondisk=sda
>>> network --bootproto dhcp
>>> rootpw root
>>> auth --enableshadow --passalgo=sha512
>>> timezone --utc America/New_York
>>> vznetcfg --net=virt_network1:eth0
>>> vztturlmap $FS_SERVER http://myrepository.com
>>> %eztmplates --cache
>>> I have tried all kinds of option combination, also deleted
>>> all the partitions, but it still can't work. Any
>>> suggestions/comments are welcomed!
>>> Best Regards,
>>> arm mailing list
>>> arm at lists.fedoraproject.org <mailto:arm at lists.fedoraproject.org>
>> arm mailing list
>> arm at lists.fedoraproject.org <mailto:arm at lists.fedoraproject.org>
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the arm