Boot from Ubuntu 16.04 live iso failed
-
Well I can give you a few suggestions here.
- For ubuntu menu, this line will not work: kernel http://${fog-ip}/tftpboot/ubuntu16/vmlinuz.efi The issue is /tftpboot is not accessible via the web server if you are using the http protocol. You need to make sure your protocol is in line with your location you place your files.
- http://${fog-ip}/fog/service/ipxe/hiren/hir.iso and anything else like http://${fog-ip}/${fog-webroot} if you place files in the FOG web gui path, plan on them getting clobbered the next update you install.
What I would suggest is this.
If you have an iso file make a new directory on the web server just for your iso files outside of the fog path. For example using the fog linux console the fog web gui path is /var/www/html/fog/… If it was me I would place my iso files in /var/www/html/iso Then you would reference those iso files on the http side with
http://${fog-ip}/iso/hirens.iso
and so on.For these lines
kernel http://${fog-ip}/tftpboot/ubuntu16/vmlinuz.efi initrd http://${fog-ip}/tftpboot/ubuntu16/initrd.lz
You have your protocol / file location mixed. You can either change the protocol to tftp or change your file locations
tftp:kernel tftp://${fog-ip}/tftpboot/ubuntu16/vmlinuz.efi initrd tftp://${fog-ip}/tftpboot/ubuntu16/initrd.lz
or http and move the files to /var/www/html/os/ubuntu16
kernel http://${fog-ip}/os/ubuntu16/vmlinuz.efi initrd http://${fog-ip}/os/ubuntu16/initrd.lz
-
Thanks george for the quick response.
So I have done what you mentioned:- I did move the hiren iso file to /var/www/html/iso/ and changed entry in the MENU to this:
:hiren initrd http://${fog-ip}/iso/hirens.iso chain memdisk iso raw || goto MENU
Now when I choose hiren I get this message:
MEMDISK: bootstrap too large to load.It was booting before at least to windows not it does not. DO I need to change something ins the “chain memdisk iso raw ||” line?
- For the Ubutu entry:
A. I changed the protocol to tftp:
:ubuntu kernel tftp://${fog-ip}/tftpboot/ubuntu16/vmlinuz.efi initrd tftp://${fog-ip}/tftpboot/ubuntu16/initrd.lz imgargs root=/dev/nfs boot=casper netboot=nfs nfsroot=${fog-ip}:${fog-ip}:/images/mounts/ubuntu16 locale=en_US.UTF-8 keyboard-configuration/layoutcode=us mirror/country=US boot || goto failed goto start
The result was that I get this error message:
vmlinuz.efi is in /tftpboot/ubuntu16/
B. I tried to use the http instead so I created: /var/www/html/os/ubuntu16 and extracted the ubuntu iso here.(after reading many posts I’m still not sure that I must use the iso file or extract it into this folder!) so here is the folder content:
Andhere is the MENU entry for it ::ubuntu kernel http://${fog-ip}/os/ubuntu16/vmlinuz.efi initrd http://${fog-ip}/os/ubuntu16/initrd.lz imgargs root=/dev/nfs boot=casper netboot=nfs nfsroot=${fog-ip}:${fog-ip}:/images/mounts/ubuntu16 locale=en_US.UTF-8 keyboard-configuration/layoutcode=us mirror/country=US boot || goto failed goto start
I still get the blank page when I choose to boot from the ubuntu entry.
What about the imgargs line? do I need to change the nfsroot path in this line? and what should it be in this case?
-
@amir3161 said in Boot from Ubuntu 16.04 live iso failed:
:ubuntu
kernel tftp://${fog-ip}/tftpboot/ubuntu16/vmlinuz.efi
initrd tftp://${fog-ip}/tftpboot/ubuntu16/initrd.lz
imgargs root=/dev/nfs boot=casper netboot=nfs nfsroot=${fog-ip}:${fog-ip}:/images/mounts/ubuntu16 locale=en_US.UTF-8 keyboard-configuration/layoutcode=us mirror/country=US
boot || goto failed
goto startDon’t forget you need to have the files web readable too. Right now they’re purely “read” but when iPXE goes to “start” them, it needs to execute too.
You could try:
chown -R www-data:www-data /var/www/html/os chmod -R +x /var/www/html/os
-
OK I did a quick search of the forums here is what I have.
For hirens by @Avaryan https://forums.fogproject.org/topic/10335/pxe-boot-adding-hiren-15-0-to-menu/2
item hiren Hiren's Boot CD 15.2 item return Return to Previous Menu choose --default return --timeout 15000 target && goto ${target} :hiren initrd http://${fog-ip}/iso/Hiren15.2.iso chain memdisk iso raw ||
For ubuntu/kubuntu by @kmstory https://forums.fogproject.org/topic/10403/boot-iso-from-ipxe-menu/5 Note that you will need to place the ubuntu extracted images in /images/ubuntu on the fog server to access them via nfs.
kernel nfs://${next-server}/images/xubuntu/casper/vmlinuz.efi initrd nfs://${next-server}/images/xubuntu/casper/initrd.lz imgargs vmlinuz.efi acpi=off root=/dev/nfs boot=casper netboot=nfs nfsroot=${next-server}:/images/xubuntu locale=en_US.UTF-8 keyboard-configuration/layoutcode=us mirror/country=US boot
Misc other iso images by @x23piracy https://forums.fogproject.org/topic/7674/problems-booting-a-winpe-based-rescuedisk-iso-over-pxe/7
-
@Tom-Elliott Thanks Tom. Did that but nothing changed!
-
How about example from iPXE themselves?
-
@george1421 Finally got it working for the Ubuntu! using the instruction you provided!
Regarding Hiren I still have problem. I have changed permissions as @Tom-Elliott suggested but for the /var/www/html/iso/hiren/ (hirens.iso is in this folder)
Now it gets stuck with this message:Loading boot sector... booting ...
Btw for the reference and in case someone needs it later I will post what I did here in another post.
-
@amir3161 said in Boot from Ubuntu 16.04 live iso failed:
http://${fog-ip}/iso/hirens.iso
Are you able to access this URL from your browser and download the ISO file (just for testing)?
-
@Sebastian-Roth good point. I just tested it. I can access the iso file from the FOG server machine : 127.0.0.1/iso/hiren/hirens.iso
but from other machines if I do : 192.168.18.28/iso/hiren/hirens.iso I get to the FOG
server management page. Seems like no matter what I type after the ip address it forwards me to the FOG login page.
I have checked the permissions for the iso folder and changed it to “-R 777” (I’m not sure if this is a good practice!) but still the same. -
@amir3161 Yeah this is a bit tricky. Newer FOG versions are pretty much taking over apache as it does rewrite/forward requests to make the FOG API work.
That said there is still room for what you want to do if you put your files straight into /var/www subdirectories (but not /var/www/html sub dirs!). For example you can have /var/www/isos/hirens.iso and should be able to download via http://${fog-ip}/isos/hirens.iso from a remote PC.
Be aware that getting a directory listing (URL http://${fog-ip}/isos/) won’t work as the rewrite engine is kicking in again. This is because DirectoryIndex is set and apache forwards your directory request to /isos/index.html internally. As this file does not exist the rewrite rules forward to the FOG API and from there to the management interface again.
-
@Sebastian-Roth One again spot on Sebastian
I can now boot from the iso but this is not the end of it yet! When it goes through the boot process I get this error :
This happens if I choose the Linux rescue environment . If I choose mini windows xp, everything works fine!
Any suggestion for this one ?
I also have a general question: How do I know how to treat a iso file to be able to boot from it on FOG? There is a wiki page for some (which I think is not updated in a while).
Is there any instructions to follow or it changes every now and then? -
@amir3161 I’m going to say re-download the ISO and do the checksum afterwards too. “Can’t read superblock” is a typical message from mounting a corrupt file system.
-
@Wayne-Workman I have checked the md5 checksum of the iso file and it matches the one on the website. Do you still think I need to re download?
-
@amir3161 Unfortunately it’s not possible to provide a general tutorial on how to include boot ISOs as every ISO is a bit different. Some may boot as ISO using the memdisk helper and others need to be extracted and need fiddling with the parameter till it works. When you extract the Hirens ISO check out the file
HBCD/menu-custom.lst
. It’s a huge list of how different systems can be booted (using GRUB instead of iPXE though). See, they are all different.I can confirm the error you posted as picture and I partly figured out what’s going wrong. But I don’t know how to fix it yet. I am sure we’ll figure it out at some point but I need to do other things now so I just report my findings so far. The picture below explains pretty much everything:
After some fiddling with this I thought I might just compare the md5 hashes of the squashfs images (one within the booted ISO and the other one I extracted from the ISO on my host system). They should match as they are the same files. Turns out md5sum can’t even read to the end of the file and fails in the booted live system.dmesg
reveals that the memdisk/mtdblock device (ISO in RAM) seems to be too small.Note to myself: Try PXE booting with pxelinux or grub to see if those show the same issue…
-
@amir3161 Are you still interested in this? I’ve lost track of this issue. Or maybe you figured it out yourself?