Branches
Comments
[»]
g4l 0.25 alpha 18 - Last before release??
by msetzerii - May 6th 2008 02:12:00
I put up a message on the 0.25 alpha 15 trying to get feedback, but only
had a few downloads of it, and some of 16 and 17 with no annoucement, but
no feedback good or bad.
Here are the changes listed.
Latest changes with 0.25
Moved development system from Fedora 5 to 8
Updated busybox to 1.10.1 (with patches to May 2)
Added telnetd program
Newer kernels
fixed missing lib for name resolution
moved dd, gzip and gunzip to busybox becuase of problem with
gunzip restore with the gunzip from Fedora 8 version.
compiled jetcat-mod with latest new system.
add bs=1M to dd commands in local menu (fix over 100% problem)
added autostart telnetd if kernel parameter telnetd=yes
added timeout 30 option to startup messages
added g4lmenu to show options
cleaned /dev directory to just console and raid subdirectories
use busybox mdev to dynamically create /dev entries
Added file list to local copy option like F: option on network
ftp://amd64gcc.dyndns.org/g4l-v0.25alpha18.iso
ftp://amd64gcc.dyndns.org/g4l-v0.25alpha18.devel.tar.gz
If no comments, will probable release it soon. I moved to using busybox
mdev to create device nods.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Unable to mount root
by tommeke7 - Apr 29th 2008 04:04:55
Hi,
I've put burned g4l just in linux with makecd on a cd-rom. But I get an
error while trying to start it.
On the menu, I pick the latest kernel version and here everything is
loading till I get this message:
RAMDISK: Compressed image found at block 0
Unable to load NLS charset utf8
List of all partitions:
0300 39082680 hda driver: ide-disk
0301 104391 hda1
0302 38957625 hda2
1600 40025916 hdc driver: ide-disk
1601 40025916 hdc1
1640 41358 hdd driver: ide-cdrom
No filesystem could mount root, tried: reiserfs ext3 ext2 msdos vfat
iso9660 ntfs romfs fuseblk
Kernel panic - not syncing: VFS: Unable to mount root fs on
unknown-block(1,0)
And here it hangs ..
Both hard disks are ext3 partitions .. And I guess the cd-rom gives the
error, but this has the standart fs given by makecd .. So what can be
wrong here?
thx,
Tom
[reply]
[top]
[»]
Re: Unable to mount root
by msetzerii - Apr 29th 2008 07:41:42
A couple of things.
First, when you use the makecd script, you must do so as root, and also
when you extact the development kit, otherwise, certain files are not
created.
./makecd create
will create the boot.iso file, which can then be burned to the cd as an
iso image.
cdrecord -v dev=/dev/hdc boot.iso
or
cdrecord -v blank=fast dev=/dev/hdc boot.iso
That assumes you cd is /dev/hdc
First is for a cd-r, second is cd-rw.
Did you modify the files to burn the image?
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Unable to mount root
by tommeke7 - Apr 29th 2008 08:30:01
I didn't use the development kid and did not modify any of the files.
The command that I used was the example in makecd itself, which makes the
iso and writes it to cd-r:
"sudo ./makecd /dev/cdwriter"
[reply]
[top]
[»]
Re: Unable to mount root
by msetzerii - Apr 29th 2008 22:09:42
What did you use to burn the iso image to the cd?
It must be burned as an ISO image and not as a file.
I've used the cdrecorder and k3b with linux, and nero with windows.
You mentioned makecd in the original message, and there is a makecd script
in the development kit, so that is why I thought you were using it. The
g4l.iso file is a cd image, and needs to be burned as such to a cd-r or
cd-rw.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Unable to mount root
by tommeke7 - Apr 30th 2008 05:56:30
Indeed, I made a mistake .. I just downloaded the tar file from
http://freshmeat.net/projects/g4l/ and it was the development file..
However I did not change anything to the script or the rest so I don't get
why it didn't work..
It doesn't matter however.. I downloaded the ISO file, burned it with k3b
and it's perfectly working! Now just testing everything a bit out :)
Thx for your responses!
Tom
[reply]
[top]
[»]
Please Help: Clone Small HD to Larger HD Using G4L V0.24
by KevinLuu - Apr 11th 2008 13:37:22
Hi,
I used G4L version v0.24 to clone 80GB HD to 120GB HD.
The Cloning process was successful.
I can used the new 120GB to boot the system. But when i checked the
capacity it only show 80GB instead of 120GB.
How can i solve this problem when using G4L to clone small HD to larger HD
so that the full capacity of the larger HD is seen after clone.
Thank you very much.
Kevin Luu
[reply]
[top]
[»]
Re: Please Help: Clone Small HD to Larger HD Using G4L V0.24
by msetzerii - Apr 11th 2008 17:13:28
The g4l doesn't do on the fly resize of the partition, but copies it as is.
You can then use programs like gparted or other programs to expand the
partition, or just create another partition with the extra space. The
exact options depend on the OS you are using. The exception to this, is if
you have done an NTFSCLONE backup of an ntfs partition. In which case you
would have had to create the partition before restoring the backup. The
same thing would then happen, in that if you did an ntfsclone backup of a
40GB partition, and restored it to a 100GB partition, it would still be a
40GB partition, but you could then run the ntfsresize program on the cd
with the partition.
ntfsresize /dev/hda1
And it would increase the size to use the whole partition.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Please Help: Clone Small HD to Larger HD Using G4L V0.24
by KevinLuu - Apr 11th 2008 21:59:49
Thank you very much for your help.
I was praticing using G4L to clone 80GB HD containing XP OS to and old
120GB HD to see how G4L works.
After discovering this issue about cloning small HD to larger HD i then
decided to use the same size hard disks for cloning (This is my first time
to do cloning. And i am pretty new to computer and Linux OS. I know XP OS
as casual user).
My real task at this time in the company is using G4L to clone a 250GB HD
containing Fedora 4 OS for back-up ( The person who handled this task had
left the company. He was the only one that knew Linux OS in the comapny).
Using the same G4L version CD, the same PC and the same hardware setup, I
repeated the same steps that i did earlier in pratice to clone 250GB HD
containing Fedora 4 Linux OS to an old 250GB HD also containing Fedora 4
Linux OS for backup purpose.
However this time, after about 0.76% completed, i saw many error messages
displayed on the screen.
I am not sure why this happened with 250GB HD containing Linux OS. So i
used Ctrl-C to stop the cloning process. This problem was repeatable with
250GB HD.
Howver i did not see when i cloned 80GB HD (contains XP OS) to 120GB HD
ealier in my pratice.
Do you why this problem only occurred with 250GB HD to 250GB HD cloning?.
Can G4L version V0.24 be used to clone HD with the size 250GB or larger?.
Again. Thank you very much.
Kevin Luu
[reply]
[top]
[»]
Re: Please Help: Clone Small HD to Larger HD Using G4L V0.24
by KevinLuu - Apr 12th 2008 07:49:53
> Hi,
>
> I used G4L version v0.24 to clone 80GB
> HD to 120GB HD.
>
> The Cloning process was successful.
>
> I can used the new 120GB to boot the
> system. But when i checked the capacity
> it only show 80GB instead of 120GB.
>
> How can i solve this problem when using
> G4L to clone small HD to larger HD so
> that the full capacity of the larger HD
> is seen after clone.
>
> Thank you very much.
>
> Kevin Luu
>
>
Thank you very much for your help.
I was praticing using G4L to clone 80GB HD containing XP OS to and old
120GB HD to see how G4L works. After discovering this issue about cloning
small HD to larger HD i then decided to use the same size hard disks for
cloning (This is my first time to do cloning. And i am pretty new to
computer and Linux OS. I know XP OS as casual user).
My real task at this time in the company is using G4L to clone a 250GB HD
containing Fedora 4 OS for back-up ( The person who handled this task had
left the company. He was the only one that knew Linux OS in the comapny).
Using the same G4L version CD, the same PC and the same hardware setup, I
repeated the same steps that i did earlier in pratice to clone 250GB HD
containing Fedora 4 Linux OS to an old 250GB HD also containing Fedora 4
Linux OS for backup purpose.
However this time, after about 0.76% completed, i saw many error messages
displayed on the screen. I am not sure why this happened with 250GB HD
containing Linux OS. So i used Ctrl-C to stop the cloning process. This
problem was repeatable with 250GB HD.
Howver i did not see when i cloned 80GB HD (contains XP OS) to 120GB HD
ealier in my pratice. Do you why this problem only occurred with 250GB HD
to 250GB HD cloning?. Can G4L version V0.24 be used to clone HD with the
size 250GB or larger?.
Thank you very much.
Kevin Luu
[reply]
[top]
[»]
Re: Please Help: Clone Small HD to Larger HD Using G4L V0.24
by msetzerii - Apr 12th 2008 09:24:44
Can you be specific on what options you are using to do the image. Is it a
network backup or local backup or click and clone?
After any failure, it is best to check the files in the /tmp directory to
see what the error is? Ctrl-Alt F2 to open a second terminal, and log in
with g4l id. Then look at the files in the /tmp directory, and see what
they might list as the error.
Also, what does /proc/partitions show on the drives? Do the have the same
number of heads and tracks per sectors. This generally wouldn't matter on
the copy, but sometimes cause problems with booting especially with
windows.
Also, have you checked the disk for hard errors, but rather than guessing,
checking the above first would be the best step. I have imaged large
drives, and have done dual 250GB drives in my main machine, and have 250,
300, and larger drives in my servers that backup with no problem.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Stuck at 0%
by Simon Bowen - Apr 11th 2008 08:20:33
Hi
I have been having problems with G4L. Heres the story. I have 20+ Asus EEE
PCs that I need to install a custom OS on.
I couldn't use the ISO on a CD since the Asus EEE doesn't have a CD drive
(and I have no USB CD Drive). So I managed to get G4L to boot from a USB
Pen Drive. The next problem was that G4L didn't have the Atheros L2
Network Drivers bundled with it.
Instead I booted the system with "System Rescue CD" (on a pen
drive), this has the correct network drivers, and then manually ran the
g4u scripts from the /rootfs folder created when the sources is untarred.
The problem I am having is that the progress does not move from 0%. G4L
creates the "img" file on my FTP server. But with nothing in. It
appears as though there are no permissions set on this folder, is this
correct? I am logging into the FTP server in G4L using the root
privileges.
Also whilst it is suppose to be making the image the hard drive light
shows no activity. Could the problem be that the Asus EEE PC uses a SS
hard drive?!
Any help would be appreciated very much!
Thanks Simon
[reply]
[top]
[»]
Use G4L clone small HD to large HD
by KevinLuu - Apr 11th 2008 13:32:47
Hi,
I used G4L version v0.24 to clone 80GB HD to 120GB HD.
The Cloning process was successful.
I can used the new 120GB to boot the system.
But when i checked the capacity it only show 80GB instead of 120GB.
How can i solve this problem when using G4L to clone small HD to larger HD
so that the full capacity of the larger HD is seen after clone.
Thank you very muck.
Kevin Luu
[reply]
[top]
[»]
Re: Stuck at 0%
by msetzerii - Apr 11th 2008 17:25:51
The problem that you are seeing is that the jetcat-mod program is not in
the path, and it is used in the process to pipe the data thru and create
the status of what is going on.
You can simple copy the jetcat-mod program to one of the directories in
your path, and then it would find it.
I have a setup that allowed for running g4l from other boot media like a
Knoppix livecd, in which you would boot from that media, and then download
the g4l programs.
ftp ftp://amd64gcc.dyndns.org/files4.gz
tar -zxvf files4.gz
./g4l
That file contains the files that g4l needs, and may not be on the media
used to boot. It places the current directory in the path, so it can find
the files. These are older versions of the files, and don't contain some
of the newer options, but for the standard imaging it is basically the
same.
Additionally, what method does you media use to boot. If it is using
syslinux, you could also just copy the bz24.x kernel or whatever you want
to use, and the ramdisk.gz file to the boot directory, and add the lines
the the syslinux.cfg file. I've done this with a flash, but in that case I
just copied the contents of the CD to the flash, and then ran syslinux on
it to make it bootable. The only other change was to copy the isolinux.cfg
to the syslinux.cfg.
Again, coping the jetcat-mod to a directory in the path would be the
quickest solution, but the others would give you access to other things as
well. One other note, I'm assuming it has the ncftp files and lzop on the
drive, and it not, you would also have to copy those programs as well.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Stuck at 0%
by Simon Bowen - Apr 14th 2008 03:33:20
Hi
Thanks for the reply. I turns out that the reason it was stuck at 0% was
that the img folder on the FTP server hadn't had its permissions set
correctly.
Now I am having problems restoring the image I have created. I am getting
errors with stdin telling me it is not an lzop file.
So I have attempted the method you suggested about downloading the file
over FTP and extract them. This initially works well. But once again the
Backup process is stuck at 0%. The "img" file on the ftp server has the
right permissions. And jetcat-mod is all there!?!?!?!
The reason I am booting into system rescue CD first is because they have
the NETWORK drivers loaded. Whereas G4L hasn't the network drivers for my
hard ware bundled with it.n
[reply]
[top]
[»]
no space left on device stdout!
by fredvr - Mar 13th 2008 04:23:03
Hi
I'm trying to use g4l to clone a fedora fc8 system disk into a backupfile
after doing the basic installation\configuration before data is added. It
has a 160Gb disk \ partition. the system is installed on a 50Gb log.vol.
(more or less doing a default setup)
the local disk I want to write the backup file to is also 160Gb with an
empty ntfs partition.
Shortly after the start of the backup process it stops and a message is
displayed:
gunzip no space left on device stdout.
The disk to clone is basically empty for about 90%.
After making the backup and compressing the backupfile will probably not
be much more than 10Gb.
How much scratch size does g4l need? (=how big must my target disk be?) or
is it not possible to store the backupfile on an NTFS partition this
way?
[reply]
[top]
[»]
Re: no space left on device stdout!
by msetzerii - Mar 13th 2008 04:53:49
Couple of things?
What version of g4l are you using? Early versions could not write to ntfs
partitions at all, so are you using the latest version or which exact
version are your using?
The ntfs partitions must have been cleanly shutdown, if the ntfs partition
was not cleanly shut down, it will fail to mount, and you will be writing
to the a very small ram drive.
What exact options did you use with the g4l?
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: no space left on device stdout!
by fredvr - Mar 13th 2008 05:37:20
> Couple of things?
> What version of g4l are you using? Early
> versions could not write to ntfs
> partitions at all, so are you using the
> latest version or which exact version
> are your using?
> The ntfs partitions must have been
> cleanly shutdown, if the ntfs partition
> was not cleanly shut down, it will fail
> to mount, and you will be writing to the
> a very small ram drive.
>
> What exact options did you use with the
> g4l?
>
thanks for your quickly reply. Using v0.24 gt3
I replaced the dest.disk with a 250Gb one which is running now and this
seems to work ok up to now. When it has finished I will try to reproduce
the error again with the other disk. You might be right about the fact
that NTFS had it's dirty flag enabled! IŽll mount it to XP 1st.
How long the running process is going to last is hard to say because of
some strange reporting values!
the progress reports at this time: 322% finished, 530832 Mb of 157066
Mb... :)
[reply]
[top]
[»]
Re: no space left on device stdout!
by fredvr - Mar 13th 2008 07:43:33
> Couple of things?
> What version of g4l are you using? Early
> versions could not write to ntfs
> partitions at all, so are you using the
> latest version or which exact version
> are your using?
> The ntfs partitions must have been
> cleanly shutdown, if the ntfs partition
> was not cleanly shut down, it will fail
> to mount, and you will be writing to the
> a very small ram drive.
>
> What exact options did you use with the
> g4l?
>
thanks for your quickly reply. Using v0.24 gt3
I replaced the dest.disk with a 250Gb one which is running now and this
seems to work ok up to now. When it has finished I will try to reproduce
the error again with the other disk. You might be right about the fact
that NTFS had it's dirty flag enabled! IŽll mount it to XP 1st.
How long the running process is going to last is hard to say because of
some strange reporting values!
the progress reports at this time: 322% finished, 530832 Mb of 157066
Mb... :)
[reply]
[top]
[»]
Re: no space left on device stdout!
by msetzerii - Mar 14th 2008 03:32:41
Can you check the contents of
/proc/partitions
It should be getting the of the partition, but the size it showed was like
150GB not 250GB, but that might have been the size of the partition.
But that doesn't explain how the number of bytes can go beyond the size of
the partition. The dd command should stop when it reaches the end of the
disk or partition, so unless it is getting information that doesn't match
the physical drive.
The /proc/partitions should give some info.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
"X - Reboot" function not working
by Frank - Feb 8th 2008 08:07:44
Hi there,
This is a nice software. Takes 20mins to backup my partition and 10 to
restore it byte by byte (45mins and 75mins for Acronis!). I can reboot
just like if nothing bad happened! :) Works like a charm with Linux as the
OS (openSUSE 64bits).
I rated this only 9 out of 10, cuz it's missing 3 functionalities and I
have one bug/problem.
Function1: When you go to command prompt from g4l, there is no way to
reboot the computer (haven't found it. Tried init, telinit,
shutdown).
Function2: Have the ability to write the images on a XFS partition.
Function3: How complicated would it be to backup only the used portion of
the filesystem (to reduce image space)? I am pretty sure there is an
answer to that somewhere if I google better.
Bug/problem: Very annoying. When I hit the "X - Reboot etc"
function from the backup panel, whether I hit X key or enter on the X
line, the screen flickers 1 tenth of a second but stays in the panel,
nothing happens, no reboot, no freeze, nothing, just a quick flicker. I
cannot reboot unless I cold or warm RESET/Power Off the machine. Then at
each reboot Linux reads the partition's filesystem journal, cuz it did not
umount as expected. I know computers and one day that unexpected reboot I
have to force will cause a problem in the integrity, though the chance is
very very minimal. If the probability exists, it will happen one day. I
have tried with all the available kernels g4l uses. Same problem.
Can you plz comment on this?
Thank you very much!!
Frank
[reply]
[top]
[»]
Re: "X - Reboot" function not working
by msetzerii - Feb 8th 2008 14:08:45
The reboot option on g4l is reboot?
If you can, boot the system from the cd, and run reboot directly from the
command line, and perhaps some message is coming up. That is all the X
option does. Not sure what that could be. That is the busybox function I
believe? I am about to release a new version, and it has busybox 1.90, so
it might be fixed in that.
ftp://amd64gcc.dyndns.org/g4l-24alpha5.iso
Latest alpha of next version.
Regarding XFS, I believe it is built into the kernel for support, but not
sure if it can mount in read/write mode.
After booting from the cd, try a mount /dev/??? /mnt/local with an XFS
partition, and see if it mounts or gives an error. If it mounts, is it in
rw or ro mode. I had to add an ntfs-3g option to get it to mount ntfs in
rw mode, perhaps something for XFS is required as well.
Thanks for the feedback.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: "X - Reboot" function not working
by msetzerii - Feb 8th 2008 15:02:34
Just to add a note. Pressing the reset with g4l should not cause any
problems, since it is running from a ram disk. Only if you have mounted a
local partition could there be an issue.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: "X - Reboot" function not working
by Frank - Feb 9th 2008 19:01:29
> Just to add a note. Pressing the reset
> with g4l should not cause any problems,
> since it is running from a ram disk.
> Only if you have mounted a local
> partition could there be an issue.
>
Mike, tnx for the very quick response. If "reboot" is the command line,
then I feel stupid :), I assume I probably did not read carefully the
instructions or I went through the list of commands (when using tab, you
can use it in prompt if I remember, I think it was with g4l) too quickly
and missed reboot. So simple I never thought about it. Sorry my bad.
Well 0.24a is perfect! I hit the X and enter and it sends the term signal
and reboot I go! Wasn't able to reproduce the problem with that
version.
As for XFS, I might give it a try, or for my next OS install I might trash
my ext3 partition and make it XFS to optimize my file usage and then mount
it manually like you mentioned. I think same thing would also apply for
JFS, but I don't use that one, I have no need to minimize my cpu usage
(which is the most important performance characteristic of JFS).
Now for pressing reset, you are totally right, it uses ramdisk. Hum, I
wonder which journal then it was constantly reading upon reboot after
using 0.23. Did not happen with 0.24a. Anyway, 0.24a works great for what
I need! :)
I'll try and have some fun copying a HFS+ partition on a MAC. :)
Tnx a lot for your answers, for my needs it deserves a 10, now! :)
Frank
[reply]
[top]
[»]
Re: "X - Reboot" function not working
by msetzerii - Feb 10th 2008 02:36:02
Not sure how the mount command works with those file systems. I was under
the impression that it could try to mount the partition and figure out the
type. The ntfs partitions were mounted correctly, but in ro mode, so I had
to add those special options. The JFS might just require using -t jfs, or
perhaps some addition support files need to be added.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Cannot boot after ClicknClone
by RK - Jan 9th 2008 07:50:08
I had done the cloning and all went thru fine. But when I tried to boot,
system just hangs at "Verifying DMI Pool Data". Worst is the
problem happens for the original disk as well.
I did the cloning using the same disk type and size but using another
machine. I really am lost at what to do. Any help would be much
appreciated.
[reply]
[top]
[»]
Re: Cannot boot after ClicknClone
by msetzerii - Jan 9th 2008 11:16:09
The clicknclone is basically a bit level copy using dd to copy from the one
device to the other. I don't see how it could have effected the original
disk, since it only reads from the source drive. Unless the copy was done
from the wrong drive.
Another possible issue might be that the original drive configuration of
the drives in the way of heads /sectors per track may be different from
the original options that were used. This might be way the original drive
is not working, if somehow this bios information has been changed.
You didn't mention the OS on the drive, but a web search seems to point to
a problem with the boot files, so it might be using fixmbr with
windows.
Again, the original drive not working doesn't make sense unless something
else has changed in the environment.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
I am in trouble
by coolguy789 - Jan 7th 2008 22:39:42
Hi everyone,
I used the g4l in order to clone my local IDE hdd to another local IDE
hdd. Both are on the same channel. I used the RAW mode, but I chose the
source as hda and the target as hda2. I assumed that it will just create
an image file on the hard disk, regardless what target hard disk I select.
It started to make the cloning, but I canceled it after few seconds since
it would take lots of time to complete. Now my system is not booting and
it is giving me the error message: "Volume group VolGroup00 not
found" "switchroot: mount failed: No such file or
directory" "Kernel Panic - not syncing: attempted to kill
init". Do you know what could happen? Are the boot files now
corrupted? Do you know if there is a way to recover from this problem? I
did the same process again, but now I chose the source as hda and the
target as hdb2. I am not sure if this will help. It is still running, I
will see in the morning, but I doubt.
Any help appreciated.
Thanks.
[reply]
[top]
[»]
Re: I am in trouble
by msetzerii - Jan 7th 2008 23:36:46
Not good. You would probable need to use the rescue cd to see if it can fix
the problem. Hopefully, it is just the boot partition that is message up,
and you can reinstall the grub.
Option A is talking about where you are going to place the image (Target)
or where you are going to read the image from (Source).
So, it appears you did the opposite of what it should be. In this case if
you wanted to make an image of hda onto hdb2, then the hdb2 would have
been the target. and hda would be what you would select as the disk or
partition to backup.
It should have failed to mount the /dev/hda, but it might have tried to
write to it directly. I haven't done a lot of modifications the the local
backup options, since I took over the project. That is why I recommend the
network options, but I think the rescue cd is the option to try and repair
the setup first.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
NTFSCLONE works just dandy, thanks.
by penguinroar - Oct 20th 2007 15:08:08
Tried NTFSCLONE out and it was a nice experience. Works like a charm, is
fast and makes small images. A 30 GB disk with 20 GB of data on it became
a 6 GB image with lzop compression. Its a really welcome addition for when
one have to make clones of ntfs partitions.
[reply]
[top]
[»]
Keyboard input two characters! Extra echo.
by Aditya Nag - Oct 9th 2007 00:38:17
Hi,
I'm trying to run G4l off a network boot on a Dell Inspiron 530s. It boots
fine, but at the command prompt, when I type "g", it inputs
"gg"!! Basically, every keypress gets registered twice.
This is a USB kb. I tried another one, same problem. I can't plug in a
PS/2 kb, since this machine does not have the ports.
The funny thing is, at the boot: prompt (the place where I can select the
kernel), the KB works fine.
This leads me to belive that it's something to do with the G4l kernel, not
the syslinux kernel.
Any ideas? Can I give any kernel parameters or something to fix this?
I have looked in the BIOS and tried changing all the USB related options..
no luck
Thanks,
Aditya
[reply]
[top]
[»]
Re: Keyboard input two characters! Extra echo.
by Aditya Nag - Oct 9th 2007 01:12:31
> Hi,
> I'm trying to run G4l off a network boot
> on a Dell Inspiron 530s. It boots fine,
> but at the command prompt, when I type
> "g", it inputs
> "gg"!! Basically, every
> keypress gets registered twice.
>
> This is a USB kb. I tried another one,
> same problem. I can't plug in a PS/2 kb,
> since this machine does not have the
> ports.
>
> The funny thing is, at the boot: prompt
> (the place where I can select the
> kernel), the KB works fine.
>
> This leads me to belive that it's
> something to do with the G4l kernel, not
> the syslinux kernel.
>
> Any ideas? Can I give any kernel
> parameters or something to fix this?
>
> I have looked in the BIOS and tried
> changing all the USB related options..
> no luck
>
> Thanks,
> Aditya
I fixed this problem. I was not booting the correct kernel. Now that I
boot bz22.2 it's fine.
But I have a NEW problem. The network card is not getting detected.
This is a Intel G33 chipset, on a Dell custom board.
I have downloaded the correct driver, from
http://downloadcenter.intel.com/confirm.aspx?httpDown=http://downloadmirror.intel.com/9180/eng/e1000-7.6.9.tar.gz&agr=N&ProductID=983&DwnldId=9180&strOSs=All&OSFullName=All%20Operating%20Systems&lang=eng
My question is simple: How do I add this driver to G4l? Do I need to
build a new kernel? Or is there an easier way?
Any help would be appreciated.
[reply]
[top]
[»]
Re: Keyboard input two characters! Extra echo.
by msetzerii - Oct 9th 2007 08:21:41
The g4l had the latest released kernel at the time, and the latest git
kernel. Sometimes the later git kernels support the latest hardware. There
have been some later kernels, and some newer git updates as well. I've
built some new kernels, but the problem with Symantec has prevented me
from working on integrating them. Do you know if the knoppix cd recognizes
the nic. There are ways to run the g4l from a knoppix cd.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Problem from Symantec
by msetzerii - Sep 28th 2007 19:40:43
Just as a heads up. I was contacted by an Attorney representing Symantec on
9/28/07 concerning the use of the word "GHOST" and references to Norton and
Symantec. I wasn't the original author that created the name of the
project, and was under the impression that name and references to Norton
and Symantec were done correctly, and that the Info screen at the
beginning of the g4l script clearly stated that it was not connected to
Norton Ghost or Symantec, and included there web site. But it appears that
is not the case.
I have sent a few messages out including one to the Software Freedom Law
Center, but have only gotten a form email back stating the limits of their
service, and to confirm that I understood it. I did confirm that I
understood, and was just looking for some guidance on what I would need to
do.
I would have no problem removing all references to Ghost, Norton and
Symantec, but really didn't want to have to change the name from G4L to
something else, but from the last email it doesn't appear that is
currently exceptionable to them.
From the latest email
We ask that you not use GHOST, SYMANTEC, NORTON or any other Symantec
trademark as the name of your program or in any other trademark sense. In
addition, changing to G4L is not sufficient. The G in the acronym stands
for GHOST so that you are still making reference to and drawing off of
Symantec's proprietary name and products. Please adopt a new name for the
program that does not draw upon or make any reference to GHOST. Ghost for
Linux and G4L should be replaced with this new name everywhere--on all web
sites, in the program itself, etc.
End of email quote
It is interesting that Ghosting on wikipedia has one of its definitions as
being disk imaging, Something like people using the Term Xerox to making a
copy even if it isn't with a Xerox machine.
The original message was also directed to sourceforge, but I have not
heard from them yet, but would expect something soon, and hopefully some
guidelines on what action needs to take place and how quickly. I did just
ask the Symantec attorney for a time line, but don't know it will be
enough to make an orderly change over, or require a shutdown of the
current project, and then a full creation of a new project with all the
changes made...
Generally, I've seen a good number of downloads each month of the various
files, but don't see much feedback of comments on the pages, but would
hate to see the project be killed by this...
Thank You.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Problem from Symantec
by penguinroar - Oct 17th 2007 12:41:59
For starter i think g4l is the best tool i have for imaging and cloning. I
have tried the others but surprisingly g4l beats them in ease of use and
stability / predictability. The speed isnt bad either compared to other
products that only clone files. Compability with partition types is one of
a kind wich makes it very useful for server cloning where no other product
can support the filesystem or partition type. Im grateful to be able to
use g4l.
My suggestion in this case would be to turn to Groklaw.net and Pamela
Jones for support. Its an OSS site thats very fluent in matters of law and
have a big following of both laymen and proffesionals that can give you
some direction.
IANAL but from what i understand its perfectly in order to reference
competing products in documetation etc. In interfaces its even more common
since its hard to say for eg. "Microsoft AD integration" without using
Microsofts trademarks. I see this letter as a simple attempt to scare. Why
g4l would infringe on anything is beyond me.
OTOH a new name could be a good thing. Something more descriptive like
Clonix, Mirrox or Linux Imaging System perhaps.
[reply]
[top]
[»]
Re: Problem from Symantec
by vint - Nov 5th 2007 06:24:01
penguinroar, I agree with you 100%
[reply]
[top]
[»]
Re: Problem from Symantec
by Rick Moen - Nov 21st 2007 15:14:01
Michael --
I have particular opinions about Symantec's demands in this matter, but
am not an attorney (but rather a computer network consultant and sysadmin
with a longtime interest in law), and so cannot lawfully advise you on
your specific legal situation, as doing so would violate the Unauthorized
Practice of Law (UPL) statutes. Attorneys, ironically, in the general
case are also barred from commenting by ethics rules and statutes, for
free or otherwise, except in the rare case where they're accredited to the
Bar in your jurisdiction and have done in-person consultation with you so
as to have done due diligence in getting all relevant facts.
So, since I cannot lawfully address your specific legal situation (and
so you'll have to guess my opinion on that), instead I'll speak in general
terms about law, something anyone is permitted to do and is (in my view)
part of citizenship. The applicable field of law is, of course,
trademark.
It's extremely common for companies to make excessive, ludicrous,
unenforceable demands on third parties about those third parties mentions
of their trademarked phrases, styles, names, etc. The reason they do this
is that the fundamental nature of trademark law makes them a little crazy
and motivates them to do it, in order to attempt to absolutely control
third-party mentions and stave off the day when their marks are
adjudicated to have become "generic" and thus no longer theirs at all.
Cory Doctorow has explained
those bizarre realities better than anyone I've seen, though I've also
maintained my own modest effort.
The fact is that trademark owners simply lack the legal right to
require you to follow their rules and/or seek their permission to mention
(or cite) the names, styles, slogans, etc., of their trademarked goods and
services. They'd really love to convince you they have those rights, for
reasons Cory explains. It's called "bluffing". They do have the right to
enjoin you from using those marks in particular ways: They can get civil
judgements against you for using their marks in ways likely to convince
their customers that your competing commercial product or service within
the same trade or industry is produced or endorsed by them -- which is
called the tort of "trademark infringement". Note that, among other
things, any 100% non-commercial third-party usage one makes of a mark
inherently never can infringe trademark.
Essentially, trademark is a (very) limited monopoly on commercial brand
impression against competing commercial offerings. It is not the
"ownership of a name" that trademark owners would like to convince you it
is.
(Not all non-commercial usage is absolutely privileged, because of the
separate tort of trademark tarnishment aka disparagement: Your
non-commercial Flash photo showing Disney's Goofy committing disgusting
sexual acts would be trademark tarnishment.)
The USA recognises the specific categories of "fair use" and
"nominative use" as types of mark usage even by commercial competitors
that is always lawful. Nominative
use is your merely naming (thus, "nominative") a product or service,
e.g., I can talk all day long about having bought an Apple iPod, and can
even tell people it sucks, without having to fear Apple suing me over
non-permitted uses of the words "Apple" and "iPod".
Fair use: Trademark holders cannot at any time bar people from using
trademark-encumbered names in their generic, non-product-specific senses.
E.g., Apple Computer cannot prevent you from writing about apples, and you
can offer macintoshes to ward off rain without needing to worry about
itinerant lawyers from Cupertino.
(Some write-ups you'll see mix up the names for those two concepts, or
use variants, but the point is that both types of usage are absolutely
lawful, regardless of what the trademark owner says, and no matter how
loudly it screams and its lawyers jump up and down.)
Anyhow, I'm absolutely aching to tell you what I think of Symantec's
attempt to push around people who don't know trademark law very well, as
most people don't -- but my doing so could be construed as UPL, so I
shouldn't.
If you visit the "ChillingEffects.org" Web site, staffed by law-school
students, you'll read interesting analyses and write-ups on trademark law.
Since they face the same UPL dilemma, their approach is to reproduce on
the Web demand letters people submit to them, yellow-highlighting
particular phrases whose meaning and context within trademark law they
then explain in detail. I.e., they cannot say "Fooware's demand letter is
full of used cow food", but they can explain what's going on in such a
fashion that you're free to draw that conclusion for yourself.
For what it's worth, I'm one of the editors of Linux Gazette, which
went through a similar situation when the corporation that then published
Linux Journal made baseless demands and legal threats purporting to
require us to change the name of our magazine on trademark grounds. We
did careful legal analysis and then politely declined. After a couple of
years of "scary" but toothless sabre-rattling, the corporation gave up and
we kept our name.
Rick Moen
rick@linuxmafia.com
(speaking only for himself)
[reply]
[top]
[»]
Re: Problem from Symantec
by msetzerii - Nov 22nd 2007 01:58:47
First, I would like to really thank you for taking the time to make this
complete set of comments taking into account the various limitation
factors.
I wish I could get some real specific answers, rather than just getting
the one side from their lawyer. Have never gotten any response from
legal@sourceforge.
The issue was discussed on Linux.com and backupcentral.com, but again.
Nothing clear, and some opinions on both sides of the issue.
I even sent emails to the government trademark office asking for some
specifics, and some guidelines on what is involved in determining if a
trademark has become generic.
The only reply was to call a long distance number.
I did remove all the references to their G word, S word and N word from
the web sites, and have also removed them from the working copy of the
main script, in which it
only appeared on one screen, where it mentioned that the program was not
related to their product, and gave their web site link if that is what
users were looking for. The
other scripts have no use of those words.
The excessive request of removing the G4L or actually just the G would be
a major problem, since it is the letter used in all the scripts, and also
the documentation uses that letter,
and then I have all the references to G4U to provide the clarification on
that issue. I just left the 3 letter names of both projects.
I did sent an email to their lawyer asking for more information, but have
not gotten any further responses back.
I had contacted the G4U author after I was contacted by them, and he had
heard nothing at that time, and don't know if he has had a similar issue,
but he had used the G word with an ing, so don't know.
Again, Thanks for your time.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Problem from Symantec
by Rick Moen - Nov 22nd 2007 02:48:49
You're certainly more than welcome, Michael. I hope those explanations
(of the general legal framework) prove useful.
(Regarding: legal@sourceforge) Hmm, legal@sourceforge is the remnant
of what was once VA Linux Systems + Andover.net, after multiple staff
turnovers, exhaustion of the IPO proceeds, and a couple of radical changes
of business focus -- and owns Freshmeat, which came to them as an Andover
property. There is or was at least one person there who's brilliant and
whom I respect highly: chief counsel Jay Seirmarco. However, their legal
department has no mission to provide legal advice to the broader open
source community, and for reasons I explained could not lawfully and
ethically give you legal advice anyway (Bar accreditation and ethics
requirements).
I hope you will find time to follow the links I suggested, which to my
best understanding accurately detail what trademark rights really are,
and, equally important, what they are not. (Again, I am not a lawyere,
cannot lawfully offer you "legal advice", and have tried very hard not to
purport to.)
There are always "some opinions on both sides of the issue" when
trademark demand letters arrive at open source projects. (The syndrome is
ongoing, which is why CentOS rather ignominiously expunged all mention of
Red Hat from its Web site, and now says only that it does rebuilds from
source code of a "major North American Enterprise Linux distribution" --
closely paraphrased -- which strikes me as a rather feeble and unfortunate
reaction on, I am told, the site webmaster's part.) The problem is that
99% of the people who opine on the subject know fsck-all about trademark
law, and at most are burbling about what they want trademark law to be.
I wish I could justify giving you advice on the legal merits of
Symantec's complaint, but you'll just have to figure that out and make up
your own mind. (I will, however, let slip that derisive laughter was
involved.)
In general, what I'd really like would be for the open source community
to show a little more backbone when faced with a 41 cent demand letter
thrown together from generic threat-letter templates by some corporation's
paralegal. Of course, "backbone" should logically require that threat
recipients take the trouble to actually research and understand the law
(or consult an attorney), which rarely happens: It's much easier to just
capitulate.
We at Linux Gazette were one of the rare exceptions. After careful
study, we said "No", and the other guy blinked, because he in fact had
been outrageously bluffing.
But most people immediately and completely cave in. And the corporate
flunkies with the rolls of 41 cent stamps know that.
Your situation may differ from ours. I would encourage you to study
what the law really is (as opposed to what net.randoms on Linux.com and
backupcentral.com, not to mention a Symantec lawyer, claim it is). It's
not really very difficult, but you'll need to decide that for yourself.
Rick Moen
rick@linuxmafia.com
[reply]
[top]
[»]
Great app
by David - Sep 20th 2007 14:04:51
Hi, I must say this app saved my day. Now for the question:
I was able to set this up in network boot and I want to automate the
installation of my image. I made it as far as the shell prompt (after the
dialogs) and was able to modify the ramdisk to call whichever command I
want.
Is there a parameter I can pass so it automatically restores the image
without going through the menu? For exemple:
g4l 192.168.0.1 image1.lzo '-u user1 -p password1' dhcp img2 restore
reboot
Crossing my fingers that you say yes.... :)
[reply]
[top]
[»]
Re: Great app
by msetzerii - Sep 20th 2007 19:31:36
G4L script is not setup to do an automatic restore, since one would have to
modify the CD image first, and if one put that in the wrong machine....
One could extract the lines of code needed.
cat g4l | grep ncftpget | grep lzop
That would show the line from the script, and you would then need to
change all the variables to the actual values. You could add this in a
front-end script. Normally, it displays a menu that feeds options to the
regular g4l script, but one could put whatever commands they want in that
script. You could just enter the modified line at the top and do something
like this.
ncftpget ....... &
xprogress
sleep 20
reboot
Note that the g4l script uses the & on the command to do the imaging, so
it runs in the background, so the progress subroutine can display the
progress. The xprogress script is one that just displays the progress. I
then usually have a short sleep to make sure all buffers are flushed, and
then the reboot.
But be very careful when testing, since doing it wrong wipes out the
destination drive.
As a note, when I am doing a large number of machines like my classroom, I
generally use udpcast to use multicast to image all the machines at one.
You can actually have it build custom CDs with options set. I have one CD
configured to send the hda image, and 7 cds that have the configuration to
restore image to hda. I boot the setup machine with the send cd, and boot
up the other 18 machines using the receive cd. After all machines at the
prompt, Press Enter and the process goes thru and is set to automatically
reboot.
So, there are options, but would take some planning to make sure you get
what you need.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
VMWare SCSI Hard Disk
by Yoway Buorn - Sep 14th 2007 11:34:38
The g4l ISO doesn't recognize my VMWare SCSI drive. Do I need to add
specific drivers to the ISO before booting?
[reply]
[top]
[»]
Re: VMWare SCSI Hard Disk
by msetzerii - Sep 14th 2007 16:07:23
G4L should generally see the hardware at the physical level, since the
vmware is not running when you boot from the g4l cd. What does the
/proc/partitions show? Also, which version of g4l are you using, and which
kernels have you tried. Sometimes the default kernel supports hardware, and
sometimes the latest git kernel does, and sometimes the older kernels work
best.
What is the controller used on the system? I build the kernels from
kernel.org, and try to include all the disk and nic controllers except for
the experimental ones.
Would probable be best to mention how the OS you run on the system sees
the drive. Perhaps what it shows in dmesg, and perhaps what happens in
dmesg after a boot from the cd. Might show an error message that would
point to a possible problem.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: VMWare SCSI Hard Disk
by flipp - Sep 25th 2007 20:37:39
> G4L should generally see the hardware at
> the physical level, since the vmware is
> not running when you boot from the g4l
> cd.
i think his issue of g4l not being able to see vmware scsi disks is
because he was booting a vmware guest machine with g4l. i was trying to
do a physical 2 virtual (p2v) move from a linux machine to a vmware linux
guest. i was able to pxeboot into g4l on the physical linux box and was
able to create a .img disk image. i then setup the vmware linux guest, set
it to pxeboot g4l, and when g4l was up it wasn't able to see any of vmware
linux guest's harddrives.
i believe it's the lsilogic and vmware scsi buslogic modules that g4l
might need. i'm going to try to use virtual ide harddrives instead of scsi
ones and see if g4l is able to recognized the disks.
http://www.rtfm-ed.co.uk/?page_id=174 does list the 2 drivers, but i
believe those are just windows drivers in the .zip files.
---
if you're interested here's the steps i followed to use g4l to p2v a linux
machine.
1. pxe boot physical machine with g4l to create machine.img
2. qemu-img convert -f raw machine.img -O vmdk machine.vmdk
3. vmware-vdiskmanager -r machine.vmdk -t 1 machine.vmdk
4. create a duplicate machine in vmware ( size of scsi:0.0 > size of
current physical drive )
5. get the mac address from the new machine.vmx file on the vmware server
[ or esx server ]
6. pxe boot physical machine with g4l to restore machine.img
at this point i was hoping to just do a restore, but after i set
everything up and clicked restore/ok it just blinked and i was still at
the menu. pop'd back to the shell and i tried to do an fdisk on /dev/sda
and /dev/hda and neither worked. i didn't poke around too much
harder.
i think i might be able to take the new vmware machine.vmx file and modify
it to match the machine.vmdk/machine-s0X.vmdk files created from g4l,
qemu-img , and vmware-vdiskmanager.
i've used g4l for a few years now for various things, thanks it's saved my
butt a few times, thanks!
---
here's what i used for the pxeboot, g4l v0.23
- kernel g4l/bz22.2
- append initrd=/install/g4l/ramdisk.gz ramdisk_size=65536
root=/dev/ram0
[reply]
[top]
[»]
Re: VMWare SCSI Hard Disk
by flipp - Sep 28th 2007 09:38:40
ha! a coworker figured it out, and sadly it's almost too simple. after step
4 (create a vm), go into the vm's settings and change the scsi controller
type from lsilogic to buslogic. pxe boot g4l, and you can now see the
disks. just finished a restore.
[reply]
[top]
[»]
Getting ready to release 0.23
by msetzerii - Aug 14th 2007 05:02:46
Haven't gotten much feedback from the alphas, so can only assume it is
working OK, but have done basic testing.
Updated the syslinux and a few other support programs
Biggest update was adding ntfs-3g support so that with the local backup
option, one can create the images onto ntfs partitions. Also, adding
NTFSCLONE backups to this local backup.
Just added the 2.6.22.2 kernel, in addition to the latest 2.6.23 git
kernel as of 8/14.
ftp://amd64gcc.dyndns.org/g4l-23alpha13.iso
ftp://amd64gcc.dyndns.org/g4l-23alpha13-develop.gz
These are about 69MB each, since they have lots of kernels. Will probable
reduce the number of kernels before the released version. But this will
probable be the last alpha, since our College semester is starting, and
will not have as much time.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Getting ready to release 0.23
by Mark J - Aug 19th 2007 22:49:58
Hi,
I am needing to be able to boot from the CD using noapic command. I am
unable to figure out how to do it. Any suggestions?
[reply]
[top]
[»]
Re: Getting ready to release 0.23
by msetzerii - Aug 19th 2007 23:40:16
I believe you use the tab key on the kernel line you want, and then are
able to add options to the kernel line. It is standard isolinux config
file using the syslinux. I don't have a g4l cd in my lab here, so can not
make sure, but am pretty sure that is the way to do it. Had some systems
in the past that required nosmp, and sometimes used the vga=ask options to
get a better screen resolution.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Getting ready to release 0.23
by Mark J - Aug 20th 2007 10:42:40
Thanks, I will try that tonight.
[reply]
[top]
[»]
blank6 "null pointer assignment" on Vista
by greyhairweenie - Aug 4th 2007 15:03:22
blank6.exe is blowing up with "null pointer assignment" on Vista
on a Compaq w/Athlon x2 +4000 cpu. Any clues/help?
[reply]
[top]
[»]
Re: blank6 "null pointer assignment" on Vista
by msetzerii - Aug 5th 2007 00:42:16
Unfortunately, I don't have access to any machine running vista. The
blank6.exe program is a simple c program that uses standard calls. But
that might not be something that vista allows.
The program changes to the root directory of the drive, so if the user
doesn't have rights to the root directory to create the files, that might
cause the problem.
Here is the code, and it is compiled with the older borland c complier.
Might be a little off since the leading spaces my not display. If not the
code is available on the web site and the cd.
ftp://fedoragcc.dyndns.org/blank6.cpp
Actually, didn't post program well, so check the link.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
USB Drive Support
by John Irey - Jul 19th 2007 16:21:34
I have been using G4L for the last several years with great success. It
has evolved into a great tool that is always in my toolkit.
Over the years, USB devices have become more prevalent, and now are an
integral part of nearly everyones daily IT life.
This being said, it would be a great advantage for G4L to backup and
restore from a USB connected device, such as a thumb drive or external HD
enclosure.
With thumb drives of 4, 8 or even 16 GB, I have more them once wished that
I could have G4L boot form such a drive and backup/restore from it.
In many cases, new PCs are being deployed without CD or DVD drives, thus
making it more difficult to initiate the software Backup/Restore
process.
Hope this comment helps.
Thanks,
John
[reply]
[top]
[»]
Re: USB Drive Support
by msetzerii - Jul 19th 2007 20:52:02
G4L should work fine for an USB device, it uses isolinux to boot from the
cd, and would just need the USB device to boot with syslinux, and then
copy the kernel file(s) and ramdrive.gz . Unfortunately, I only have
access to two USB sticks, and I've never been able to get them to boot on
any of the machines I have access to. I've even tried other methods to get
it to boot from the USB, and it never works, so don't know if it is the USB
sticks I have or the motherboard or some setting I'm missing.
I've done a similar thing by just adding lines to the grub menu on
machines, and it loads the g4l just like any other kernel. Just add the
lines or the kernel and ramdrive.gz.
I believer there was one or more users that have been able to do it. Do
you have a bootable USB device that uses syslinux already to boot?
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: USB Drive Support
by John Irey - Jul 23rd 2007 07:29:45
Thanks for the update on the USB drive.
I have setup a USB drive that is bootable, but your tips should help get
the G4L loaded on it correctly.
I think that we could make this a little easier for some users though. If
there ware a menu selection that would run an install of G4L on a remote
drive. Then a user could boot from the CD and install G4L on any type of
media.
The slick thing about this is that it could be any type of media. an SD
card, USB drive, Local HD partition...
Interestingly enough, if you had G4L on the HD partition, you could boot
to it and restore other partitions. This works great for training rooms
and systems that are reloaded more frequently.
Coupling this media with a local image makes for an extremely fast
re-image process.
Also, to address the backup side of things, I was thinking that it would
be a good idea to be able to select the backup location, since network and
FTP is not always available.
For example, I have several PC that are in rolling presentation enclosures
not on the network. In this situation, I ether have to roll the whole
enclosure to a network connection or remove the unit from it to make the
backup. Nether of these options is all that fun.
If you do decide to incorporate this feature, I would also suggest that
you have a check to insure that users will not backup an image onto
itself. There are always those 3am stupidity moments that we have all
had.
Thanks again for your quick response and a great program.
John
[reply]
[top]
[»]
Re: USB Drive Support
by msetzerii - Jul 23rd 2007 09:13:32
I agree that it would be nice to come up with a simple system. Generally
the CD is simple process, but there have been a few users that have
systems without CD drives.
I've added to my grub menu by just putting a kernel, and the ramdrive.gz
file in the /boot directory. Probable the same can be down with lilo.
Setting it up on a Fat32 or hard disk with Fat32 and syslinux would be
easy, but the problem would be systems with just NTFS.
What process did you use to make your usb bootable? Did you use the usb
zip option.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: USB Drive Support
by John Irey - Jul 25th 2007 22:01:40
Try this process at the following link:
http://wiki.fdos.org/Installation/BootDiskCreateUSB
It worked for me
Thanks,
John
[reply]
[top]
[»]
Re: USB Drive Support
by msetzerii - Jul 26th 2007 01:56:34
Thanks for the link. I looked at a few of them, and some of those look like
things I had tried in the past, but it might be one step that was missing.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: USB Drive Support
by msetzerii - Jul 27th 2007 20:17:23
I was able to get my usb device to work using the syslinux usb options, but
had to change some things.
First, my bios didn't have options for usb-hdd, but would see the usb as a
rmt-fdd (I believe).
So, I used the following.
mkdiskimage -1 /dev/sda 0 64 32
syslinux /dev/sda1
Then I was able to mount the /dev/sda1, and then copied the contents of
the cdrom to /dev/sda1.
Then I copied isolinux.cfg to syslinux.cfg
That worked to boot up, but the longer kernel names didn't work, so I had
to rename the image name and edit the syslinux.cfg with nano to be able to
select the later image names.
Also, discovered I had to change the USB settings to accept Legacy
mouse/keyboard or it wouldn't work.
So, if one can get the syslinux installed on a device, and get it to work
with the computer bios, it should only be a matter of copying the
isolinux.cfg to syslinux.cfg.
In my case, I renamed the bzImage22.1 to bz22.1 and it worked fine.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Looking to add NTFS Write capability to Local
by msetzerii - Jul 16th 2007 22:01:37
I'm working on adding write access to NTFS partitions, and have just done
some very basics. I added ntfs-3g, fusemount, libntfs-3g... and
libfuse..... to the g4l. Then added these lines under the local backup
option.
With the regular mount, it can mount the ntfs partition, but it is only in
read mode, even thou it shows rw. The following lines check to see if the
regular mount was able to mount the volume, and it so, is it an ntfs one.
If so, it then unmounts the partition, and remounts it using the
ntfs-3g.
I have only do one test on it, and it was able to do the mount and make a
backup on my system. Worked with the 22.1 kernel, but didn't with the
default kernels, so not sure what the requirements would be, and if there
are any problems. I copied the 4 files from my Fedora Core 5 system.
else
testntfs=`cat /etc/mtab | grep /dev/$localdrive | cut -f 3 -d\ `
if [ "$testntfs" = "ntfs" ]; then
umount /mnt/local
ntfs-3g /dev/$localdrive /mnt/local >/tmp/mount 2>&1
fi
As an additional note: Didn't get the greatest speed, being that I got an
overall speed of about 11MB/second, whereas I usually get 30MB/second
using an FTP backup. But might be in the dd line options, or that this is
a 1.8Ghz machine, compared my classroom 2.8Ghz machines.
If this does work, am also looking at adding the ntfsclone option of
backup to the local copy. Didn't make a lot of sense to do it, if one
could only write to fat32 partition.
Anyone see any problems with this. Probable change the default kernel to
the 22.1, since they have seemed to be stable.
Thanks.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Looking to add NTFS Write capability to Local
by msetzerii - Jul 26th 2007 02:05:27
Just as a note. I have made an updated version that adds the NTFS write
capability local copy. That seems to work just fine. Also, added NTFSclone
to the local backup options.
I've tried doing an NTFSclone backup of my XP partition onto the FAT32
partition. The NTFSCLONE went thru fine, and created a split image, since
its size was greater than 4GB. But for some reason at the end of the
creation, it stop responding after the Sync at the end. But I was able to
do a restore of the image files with no errors, so whatever caused it to
stop occurred after it successfully finished making the image. Did a
second test, and it got past the sync, but the produced a segment fault on
the date command that it runs at the end to track the time. Again, the
restore from this backup also worked.
Then I hooked an USB drive to the system that has ext3 format, and was
able to both a single NTFSCLONE backup to a single file, and then one
using the split options. Neither had the problem.
Have placed a message on the NTFSCLONE web site, to see if there might be
somthing that they are aware of, and possible a fix.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Looking to add NTFS Write capability to Local
by msetzerii - Jul 27th 2007 20:27:35
I seem to have come up with a solution to the problem of splitting an image
onto a FAT32 partition, but don't know why it makes a difference.
Originally, the split option always used a 2000m split size. In doing a
test, I opened two other terminal windows, and ran top in one, and was
using ls in another to monitor the size of the image files being created.
At first, I could switch to the third terminal window, and do an ls, and
see the file getting bigger, but eventually it would stop responding. The
top on terminal 2 was still running fine, and so was the g4l, but ending
top would result in a freeze as well.
As a test, I modified the split size for the lzop option, and set it to
1000m, and did another image. This time it went thru all the way with no
errors. Then tried it with 1800m, and it failed again. So, figure I'll
stick with the 1000m. Don't know why this would be a problem. The 2000m
worked fine with an ext3 partition. So, it seems to get the job done, but
would really like to know what is causing the problem.
The syslinux 3.52 is suppose to be released shortly, and I have a few
other little changes I need to make. Want to add the Special Option on the
Local menu to set the auto reboot option, and have already updated the
latest udpcast files.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
md5 sums
by snaildarter - Jul 12th 2007 19:47:35
I had the idea to verify the integrity of my images by first making an
image (with g4l), then immediately making another identical image (with
g4l), then comparing md5 sums of the 2 image files, which should match.
Imagine my horror when they did not.
I did a line compare and discovered minor differences around line 10 and
line 15 in the images. I have also discovered that this is consistent
behavior for g4l. I have a few questions:
1) Why are the images different? Does g4l put a timestamp or something
like that in there?
2) Is there a way to turn that off so that I can do my integrity checking
with md5 sums?
3) If not, is there ANY way to check the integrity of image files?
And by the way, thank you for such a wonderful program!
[reply]
[top]
[»]
Re: md5 sums
by msetzerii - Jul 12th 2007 20:15:44
The image files are lzop or gzip compressed files, so the way I check the
files is by using the compression programs test.
lzop -t image.lzo
That then comes back with an OK if the file passes the test.
I would guess that making an image with none for the compression would not
add anything to the size of the image, but would make a huge file.
I did once have a system that created image files that later failed the
lzop test on the image. Turned out to be a bad memory module. Ran memtest
on system, and it went thru 7 test with no errors, but test 8 found
errors. Replace the bad module, and have had all test pass since.
Now after doing the compression programs test and having it pass, one
could create an md5sum file.
As you said, it is probable some kind of time stamp made by the
compression program, but don't know if it would be a fixed size that one
could somehow skip over. Also, what would one compare it to, since an
md5sum of the disk would not be the same since the image file is
compressed.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: md5 sums
by snaildarter - Jul 13th 2007 06:08:23
Yes, you are exactly right. I checked into it further. LZOP does indeed
store a time-stamp, which makes the md5sum method impossible. Using the
md5sum method on uncompressed images does work, however.
But since LZOP generates and stores an Adler-32 checksum on the fly, then
your method of integrity check using "lzop -t ..." is a very good one, and
is a better solution than my md5sum idea anyway!
Thank you again!
[reply]
[top]
[»]
Production -> Dev Server?
by Tom - Jul 5th 2007 23:10:16
This project looks really interesting to me because I have a customer
looking to take his production Oracle server and make a replica of it into
a Dev box.
I have tried a couple ways of making this happen, all of which are
failing.
Both machines are Sun servers, but the Dev box is only about half the
hardware of the Production machine. It has more hard drive space
though.
Would it be possible to image the Production machine (safely?) and then
drop the image onto the Dev box? Then I imagine I would just have to
change the IP Address, Hostname and such?
How would the partitions come out? The same as the production
machine?
Sorry for all the questions. :-\
Thanks,
Tom
--
Tom Wilson
Tridium Technologies, LLC
http://www.tridium-tech.com/
[reply]
[top]
[»]
Re: Production -> Dev Server?
by msetzerii - Jul 6th 2007 05:33:40
The problem is that I'm not familiar with Sun hardware. The G4L is a linux
based boot cd that works on X86 and 64bit Intel/AMD Cpus. Not sure if it
will work on Sun Systems.
If it does boot from the g4l linux cd, then it should be able to make an
image of the hard disk to an ftp server, and then restore the image to
another system.
If the regular g4l cd doesn't boot on the Sun hardware, there might be
versions of linux that do boot on Sun Hardware. The g4l script is basic
shell script, and the support programs could be gotten with that version
or recompiled source code.
The g4l is a gui to basically dd and ftp with compression. Might be
simpler since this seems to be a one time thing to just do it
directly.
Say you boot from some kind of live boot cd on the Sun system and open a
terminal window.
Here is a line from the script that makes the image
(dd bs=1M if=$disk 2>/dev/null | jetcat-mod -f 5000 -p $readsize
2>/tmp/progress.out |lzop -c - | ncftpput -o "useHELP_SITE=0" -m
$useridpass -c -d /tmp/out $server "$ftppath/$netimagename"
2>>/tmp/ncftpstatus.out)
Now that line uses the information passed from the script, and also
supports the gui progress bar via teh jetcat-mod program.
dd bs=1M if=/dev/hda 2>/dev/null | lzop -c - | ncftpput -m -u userid -p
password -c -d /tmp/out serverip img/serverimage.lzo
So, there are various options.
A final option would be to place the drive into a machine that could boot
from the g4l cd, make a backup image. Then return it to the machine, and
then restore the image to another hard disk in the machine, and then place
that one in the Dev machine.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Production -> Dev Server?
by msetzerii - Jul 6th 2007 05:33:52
The problem is that I'm not familiar with Sun hardware. The G4L is a linux
based boot cd that works on X86 and 64bit Intel/AMD Cpus. Not sure if it
will work on Sun Systems.
If it does boot from the g4l linux cd, then it should be able to make an
image of the hard disk to an ftp server, and then restore the image to
another system.
If the regular g4l cd doesn't boot on the Sun hardware, there might be
versions of linux that do boot on Sun Hardware. The g4l script is basic
shell script, and the support programs could be gotten with that version
or recompiled source code.
The g4l is a gui to basically dd and ftp with compression. Might be
simpler since this seems to be a one time thing to just do it
directly.
Say you boot from some kind of live boot cd on the Sun system and open a
terminal window.
Here is a line from the script that makes the image
(dd bs=1M if=$disk 2>/dev/null | jetcat-mod -f 5000 -p $readsize
2>/tmp/progress.out |lzop -c - | ncftpput -o "useHELP_SITE=0" -m
$useridpass -c -d /tmp/out $server "$ftppath/$netimagename"
2>>/tmp/ncftpstatus.out)
Now that line uses the information passed from the script, and also
supports the gui progress bar via teh jetcat-mod program.
dd bs=1M if=/dev/hda 2>/dev/null | lzop -c - | ncftpput -m -u userid -p
password -c -d /tmp/out serverip img/serverimage.lzo
So, there are various options.
A final option would be to place the drive into a machine that could boot
from the g4l cd, make a backup image. Then return it to the machine, and
then restore the image to another hard disk in the machine, and then place
that one in the Dev machine.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
need help
by deifal - Jul 4th 2007 01:20:06
i need some help!
i have some notebooks and i want to make images and store them on a serve
and when i log on to the network with the notebooks, the notebooks should
automatically be updated with the images from the server.
can you help me and tell me how i can manage this?
[reply]
[top]
[»]
Re: need help
by msetzerii - Jul 5th 2007 00:26:43
One would have to know more details of exactly what you want to do.
The first step would be to find if your hardware works with g4l. Most
systems will work with one of the kernels, but there have been some that
don't, and they require using another boot cd.
If it does successfully boot from the cd, you could then create an image
of the notebook on a server running an ftp server. This system can have
any OS, but note the windows ftp server has problems with both g4l and
g4u, but Filezilla seems to work fine as an ftp server for windows. The
image of the hard disk could then later be restored to that notebook or
other notebooks with the same size or larger disk. You also didn't
mention which OS is on the machines, if it is WIndows, then the NTFSCLONE
program that is included on the g4l is a faster way to restore the NTFS
partition.
The automatic part is going to depend on what you mean by automatic.
Inserting the g4l cd and booting from that will bring up the g4l, put the
user will have to run the program. This can be further automated by using
a front-end script, that provides options to select that can provide most
of the options. Some users have even used the development kit, and created
custom cds to do this. But one most be very careful, since when it
restores, it wipes out everything that is currently on the hard disk. It
is a true disk image.
Years ago, I had systems that would boot via WOL at 6:30am each day, they
would then use Norton Ghost to reimage the C: partition from the image
file on the hidden D: partition. This only took about 10 minutes, since
the local drive only had Windows, the apps where all run from the server.
This had a program in the autoexec that would check the date of the
program, and if it wasn't a preset date or todays date, it would then do
the reimage. If it was the preset date, it would know that it had just
imaged the disk, and would then set the date of the file to todays date,
and boot windows normall.This way on additional reboots,it would not
reimage the machine, but booting on the next day would result in another
reimage.
With boot from the Grub menu, I have even put g4l on that as a boot time
option, and some have setup PXE boot options.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Modify g4l Ramdisk
by Thomas Burnham - Jul 2nd 2007 07:40:55
Greetings;
Curious to know if there is an easy method of editing some of the options
in the g4l ramdisk from Windows (Cygwin perhaps?)
The reason I ask is, I have g4l on a DVD with my image (2.6gb). The
graphical interface will not detect the fact that there is a DVD in
/dev/hdc, so I have to manually mount that device, then manually run dd
(cd /mnt/dvd; cat hp5700.img | lzop -d -c - | dd of=/dev/hda)
I'd like to automate this process, simply edit some of the startup
options, however Cygwin errors out with Invalid Argument when I attempt to
mount the ramdisk.
Any feed back on an easy method of updating this?
Thanks;
Thomas
[reply]
[top]
[»]
Re: Modify g4l Ramdisk
by msetzerii - Jul 2nd 2007 08:24:13
The ramdisk file is not just a ramdisk, it is the linux system including
the kernel and support files. The syslinux loads the ramdisk into memory,
and then passes control the the kernel you select, and then mounts the /
of the ram disk with the other directories.
I believe you would have to use the development kit, and create a custom
ISO image. You would need to extract the development kit as root, and then
you could copy your image file to the isolinux sub directory. You would
probable have to modify the makecd script to create a DVD script instead
of a CD.
| 
