Branches
Releases
|
Version
|
Focus
|
Date
|
|
0.27
|
N/A |
20-Sep-2008 06:17 |
|
0.26
|
N/A |
10-Jul-2008 11:25 |
|
0.25
|
Minor feature enhancements |
14-May-2008 07:15 |
|
0.24
|
Minor feature enhancements |
04-Mar-2008 08:59 |
|
0.23
|
Major feature enhancements |
17-Aug-2007 13:18 |
|
0.22
|
Major feature enhancements |
23-Feb-2007 07:38 |
|
0.21
|
Major feature enhancements |
17-Jun-2006 07:52 |
|
0.20
|
Minor feature enhancements |
22-May-2006 11:32 |
|
0.19
|
Minor feature enhancements |
30-Mar-2006 14:57 |
|
0.18
|
Minor feature enhancements |
26-Jan-2006 12:52 |
Comments
[»]
g4l standalone
by Mrs. Karen Debbie - Oct 23rd 2008 00:00:49
Hello
I've been looking for tools to create backups of partitions, most of them
have ntfs file system.
I try to gather couple tools (for partitioning, backup, copying etc) on a
usb stick and sysresccd was the right thing. But I'm not satisfied with
partimage. Well, lucky me I've found g4l. I was able to launch 0.22 version
under sysresccd, but I hadn't luck creating backup due to jetcat-mod, and I
assume with paths to g4l files.
As for now I see that only way to use latest versions of g4l is to burn
.iso file on CD, but would you provide latest stand alone g4l files to use
with other linux distros ?
btw: is g4l able to create an img from from 1 partition (f.e. sda1) and
save it to other partition in the same disk(f.e. sda5) both having
ntfs.
ps: I've found 2 archives: files4 & files5, but those are 2 years old,
and ... I'm kinda scared to use them.
[reply]
[top]
[»]
Re: g4l standalone
by msetzerii - Oct 23rd 2008 02:59:05
ftp://amd64gcc.dyndns.org/finnix.gz
Is a new one that I created for finnix cd.
I tired it on knoppix like the old one, but the libraries that it has are
not the same as the programs. The older one was all statically compiled,
the later versions use the libraries. I found that the finnix had the
latest kernels, and most of the programs already on the cd, so the finnix
only needs to be there.
The jetcat-mod program needs to be in the path, or it needs to be the /
directory for the cleandrive. The filesx.gz had a modification that would
add the current directory to the path, so it would find those files.
There is also a cd that has g4l added, but I have just looked at it a
little.
http://www.tux.org/pub/people/kent-robotti/looplinux/rip/
As for making an image on the same drive this is very possible. Another
teacher has a lab that is only XP, and they have been having virus
infections that gets past N antivirus from flash disks. The disk were 80GB
but only using about 12GB. I resized the disk with gparted to 40GB for the
XP, and created an ext2 partition. Then created the image of the XP onto
the ext2 partition. This way the image is on a partition that can not
normally be seen by windows. The machines can then be booted from a g4l cd,
and the xp can be restored in about 6 minutes.
The 2nd partition could be ntfs, but then it could be seen by windows, and
viruses might effect it.
In my lab, I actually place a copy of the XP image on the linux
paritition, and can restored from the linux directly.
The g4l is image can be copied from the cd to a flash, and syslinux to
boot it from a flash. Also, the g4l can be added to the regular grub menu.
I've also heard there is grubdos, but haven't used it.
Not sure I answered everything, but there are some ideals.
Additionally, the syslinux has recently added a hybrd script that modifies
the iso image, so it can be burned to a flash just like the cd image, but I
haven't tried it yet.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: g4l standalone
by Mrs. Karen Debbie - Oct 23rd 2008 12:30:04
tbh I expected a fast reply, but not that fast :)
More than 1 distro on a usb stick cannot be used due to obvious reasons,
so I've chosen sysresccd, and it works pretty nicely with its tools
(thinked as well about usin DSL distro) .
That's why I wanted to add g4l to it (partimage isn't exactly what I was
looking for, and deveice-image is too much user unfriendly), but I don't
know how to modify img file of sysresccd and add g4l to be a native part of
system.
I've found once a solution how to save changes made in a under sysresccd
session but I can't find it anymore.
So I just copied old g4l stand alone files (those script files available
at http://sourceforge.net/projects/g4l) and tried to run it.
It worked, untill I hit OK to begin creating img of partition, then I got:
error line 1551 something with jetcat, and application hanged (as I've
noticed from reading comments it was a prob with progress bar, but there
was jetcat file in g4l catalog).
I'll try with finnix.gz (syres uses kernel v2.6.25-r4) within couple days.
I so hope they'll work without problems and I'll be able to drop my
dos_tool pendrive and advance to use linux one :)
But dependencies from system libraries are a pain in the a**, so I
slightly doubt It'll work since 1st hit.
As it goes to ntfs partitions: I just don't want to shorten users hdd
capacity, so I just create img files and store them on the last partition
on their disks. Viruses doesn't have a chance after recreating system from
img file, cause under linux distro (+ntfs-3g) I'm able to delete any
suspicious files on other ntfs partitions.
If not, I can always copy img files into flash stick, reformat&repartition
whole hdd disk and restore system from img again.
btw: every day I become convinced that windows environment fools me with
every advanced thing I want to do. From the other hand I'm as well fed up
with searching for solutions again and again needed to do smth under linux.
Good that there is this small percent of good people :)
[reply]
[top]
[»]
Re: g4l standalone
by msetzerii - Oct 23rd 2008 13:04:47
More than 1 distro on a usb stick cannot be used due to obvious reasons, so
I've chosen sysresccd, and it works pretty nicely with its tools (thinked
as well about usin DSL distro) .
I'm not clear on what the obvious reasons are? If the flash uses syslinux
or grub the g4l could easily be added. I'm not sysresccd, so not sure what
it uses. The g4l runs from a ram disk, so only needs the kernel and 64M of
Ram.
That's why I wanted to add g4l to it (partimage isn't exactly what I was
looking for, and deveice-image is too much user unfriendly), but I don't
know how to modify img file of sysresccd and add g4l to be a native part of
system. I've found once a solution how to save changes made in a under
sysresccd session but I can't find it anymore.
If the sysresccd has the files, other files, you could probable download
the finnix.gz file I mentioned, and see if the programs run. I don't know
if it has ntfsclone programs, so if not you would also have to down load
them.
So I just copied old g4l stand alone files (those script files available
at (http://sourceforge.net/projects/g4l) and
tried to run it. It worked, untill I hit OK to begin
creating img of partition, then I got:
error line 1551 something with jetcat,
and application hanged (as I've noticed
from reading comments it was a prob with
progress bar, but there was jetcat file
in g4l catalog).
The program is jetcat-mod, and it would need to be in one of the path
directories.
I'll try with finnix.gz (syres uses kernel v2.6.25-r4) within couple days.
I so hope they'll work without problems and I'll be able to drop my
dos_tool pendrive and advance to use linux one
:)
Please let me know how it works out.
I don't know if you use an iso file to create the flash, but if so,
isomaster can let you make modifications.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: g4l standalone
by Mrs. Karen Debbie - Oct 29th 2008 16:43:37
> I'm not clear on what the obvious
> reasons are? If the flash uses syslinux
> or grub the g4l could easily be added.
I completely forgot about grub (thus I was pointing of having only 1
active partition per hdd).
It makes sense now. But from other hand, having more than 2 partitions on
USB stick isn't much handy.
> Please let me know how it works out.
Ok, I think it works with syresscd without merging it with .iso.
But there were couple rasps in the whole process.
1. before launching g4l I have to create dir "local" in /mnt or error with
paths shows up, and added path to g4l directory using:
path=$path:/mnt/cdrom/g4l (cdrom is default mount point of usb stick
partition with sysresccd distro)
2. g4l has hard coded paths to everything so I had to do a small
cheat
3. since auto mounted ntfs partition is located in /mnt/local, I had to
mount another partition in /mnt/local/any_directory using ntfs-3g
4. I used options:
- ntfs clone backup
- path: any_directory/img_gile.img
- gzip compression (file not included in finnix.gz)
and hit
Whole process started in command line (no progress bar in g4l window), so
I don't know whatfor if jetcat-mod was required.
I'm quite happy that no errors popped up when creation of partition image
was processed.
Conclusions:
- g4l menu is complicated too much
raw mode / file mode -> which should I choose to backup a partition; which
is better for whole disk etc.?
- in raw mode -> local:
paths aren't saved, I don't know what's the difference between
backup/restore and ntfsclone backup/restore modes. I mean: am I able to
create img of ntfs partition using normal backup option; or is it intended
only for linux type partitions ?
btw. simplifying whole img creation process would be a good thing.
Especially if you make usage of ntfs3g library while choosing target
partition.
I don't know how does it look using g4l.iso, but using g4l as an external
program is too difficult comparing to partimage software.
> I don't know if you use an iso file to
> create the flash, but if so, isomaster
> can let you make modifications.
I'll try to mess with sysresccd.iso and g4l asap.
(read asap as "in this year maybe;)
Thx for helping :)
[reply]
[top]
[»]
v0.25 and up 515% slower
by Frank - Oct 13th 2008 18:34:18
Hi,
What changed between 0.24 and 0.25+ so that it takes 515% more time to DD?
Is it cuz DD moved to busybox?
I do RAW mode, Local files, DD Partition 1 ext3 to Partition 2 ext3, do
that every week, 15gigs, takes 8mins40sec with 0.24 and over 49mins with
0.25+.
I tried 0.27, 0.26 and 0.25, all the same. Went back to 0.24 and all is
super quick.
tnx
Frank
[reply]
[top]
[»]
Re: v0.25 and up 515% slower
by Frank - Oct 13th 2008 19:41:04
> Hi,
>
> What changed between 0.24 and 0.25+ so
> that it takes 515% more time to DD? Is
> it cuz DD moved to busybox?
>
> I do RAW mode, Local files, DD Partition
> 1 ext3 to Partition 2 ext3, do that
> every week, 15gigs, takes 8mins40sec
> with 0.24 and over 49mins with 0.25+.
>
> I tried 0.27, 0.26 and 0.25, all the
> same. Went back to 0.24 and all is super
> quick.
>
> tnx
> Frank
It seems I gain 22megs out of 2.693gigs. That's 0.82%. If the gzip
compression has been increased between 0.24 and 0.25+, well to gain 0.82%
in size at the price of 515% more in time is quite not good. :)
tnx
[reply]
[top]
[»]
Re: v0.25 and up 515% slower
by msetzerii - Oct 14th 2008 03:08:31
Interesting??
I don't believe it is a dd issue, since I do images every week at least.
More likely, it was switching from the full gzip to gzip in busybox, but
not clear on that. I generally use lzop, since it is about twice as fast as
gzip in making images, though they are about 10% or so larger.
On my systems, lzop makes an image in about 40 minutes, whereas gzip took
just under 2 hours. Restore time was very close, so it is creation process
that gzip does more, and also, the cpu load on the compression process is
higher. Gzip usually is in the 90% range, where lzop is in the 30% range.
I would assume that busybox must have done something with gzip to reduce
size, but may greatly increased time for larger processing. If you could
boot from the newer g4l cd, and replace the busybox link to gzip and
gunzip, and put the regular program to confirm that this fixes the problem.
It would be easy to switch back to the external programs, and it isn't a
great deal of extra space. It was just one less thing to do, but if it make
this big of a difference, it must be changed back.
This is the first I've heard of this, so perhaps others didn't see the
problem, or are using the lzop...
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: v0.25 and up 515% slower
by msetzerii - Oct 14th 2008 04:19:38
Did a quick test just on gzip, busybox gzip, and lzop on a 2,810,716,160
byte tar file of compressed music files.
Regular gzip: 10m8.119s
Busybox gzip: 12m47.709s 126.24% of gzip
lzop: 4m49.499s 47.61% of gzip
So, busybox gzip in this case appears to be about 26.24% slower than gzip
on this file. Not clear how you get 500%?? My test machine is an AMD
3000
From the size of compressed file:
gzip: 2,739,047,074
bb gzip: 2,753,384,110 0.51% increase
lzop: 2,788,988,956 1.78% increase
Now this is a tar with fully compressed files, so they don't really
reflect a disk image, which is a mix of compresses and uncompressed
files.
I'll try to make an alpha 5 over the next day or so of 0.28, and put the
separate dd, gzip, and gunzip on it. It is mid semester, so things are a
little busy, but I should be able to make the image, and get it
uploaded.
The latest version that doesn't yet have this change is:
ftp://amd64gcc.dyndns.org/g4l-v0.28alpha4.iso
When I do get it done, and alpha5 will be there.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: v0.25 and up 515% slower
by Frank - Oct 14th 2008 21:29:22
Hey tnx, I like your replies. That's a great thing about this project, not
only what's on CD works like a charm and is super reliable, but you also
get quick answers on questions. :)
I tested lzop 10 months ago and the difference of time vs size was better
for gzip, but I changed partition size since, I will give it another try.
I will certainly test the alpha5 too, once available. And/or try dd from
command line after booting with g4l.
I can tell though my tests between gzip and bb gzip are 100% consistent. I
always get the same times, give or take a few seconds as I don't remember
exactly.
Partition size 14,683,410 (root partition, sda1)
Used size 6,167,032 (give or take 50megs)
8mins40sec for gzip
49mins40sec for bb gzip
The first progress bar update with gzip shows between 20 to 33megs/sec.
The first progress bar update with bb gzip shows under 5megs/sec. Stays
around 3-5megs/sec throughout the backup (might increase when backing up
unsused blocks).
tnx
[reply]
[top]
[»]
Re: v0.25 and up 515% slower
by msetzerii - Oct 15th 2008 00:13:49
Just to note: The alpha5 was up about an hour after my last message. It
only took about 5 minutes to modify and build the new image, but about
25min to upload from my home machine to Office server for the iso, and then
the same for the dev kit.
ftp://amd64gcc.dyndns.org/g4l-v0.28alpha5.iso
ftp://amd64gcc.dyndns.org/g4l-v0.28a5.devel.tar.gz
As for lzop, In past testing, I found it to be about twice as fast, and
with a 10% larger size, so to me, the extra space wasn't an issue. But that
does include clearing out the unused sectors before doing the backup. Once
did a backup of a new Fedora install, and it was 12.5G on an 80GB drive.
Then cleared the unused space and redid the backup, and it was only
2GB.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: v0.25 and up 515% slower
by Frank - Oct 15th 2008 01:36:20
I tried lzop and I get twice as fast like you but with 26% bigger file on a
13.8gigs partition with 7.76gigs unused. Since to me there is no difference
between 8mins and 4mins, the extra 26% I save from gzip is interesting.
I never found out how to not dd the unused space and reduce image file. I
am unsure if when I restore, it will not take into account the unused space
and leave it unformatted/unrecognized, so I always backup full.
I have d/led alpha5 and will give it a try soon. Tnx for the help.
[reply]
[top]
[»]
Re: v0.25 and up 515% slower
by msetzerii - Oct 15th 2008 11:03:39
The 26% is a lot higher than I have ever seen, so some questions come to
mind. Had you cleared the unused space. DD does copy all sectors, but using
the cleandrive or cleandrive5 or cleandrive6 scripts to put Nulls in all
these sectors will speed up both lzop and gzip. In both cases, these will
compress to almost nothing. Unused sectors with random data take longer to
compress, and gzip does a better job. Also, what was the cpu on the system.
My machines are generally 2.8Ghz P4s with Hyperthreading. With 100Mb nics.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: v0.25 and up 515% slower
by Frank - Oct 16th 2008 00:00:49
No I did not clean the unused sectors, I must admit I don't know how.
Cleandrive is in the g4l image? Yeah I see, don't know which one to use
between 5 and 6, will try 6. I am sorry but where do I run cleandrive? I
tried after booting g4l, but could not locate cleandrive, Special Commands
function does not locate either and in the tools menu you removed the
zero-tool function and say better to use cleandrive, but how? Do I need to
use it from the OS when partitions are mounted?
I have a Quad core 2.8ghz with 100mg nic.
Ok now for the bb gzip test, it did the trick. Took 9mins02sec, slightly
slower than 8mins40sec with 0.24, not sure why, but it seems putting gzip
into bb really made a huge difference for me. I will try cleandrive maybe I
won't see the difference in time anymore, like you, and will use lzop.
It's been since January 2008 that I have started using this HDD, so no
cleandrive since then.
tnx for the support!
[reply]
[top]
[»]
Re: v0.25 and up 515% slower
by msetzerii - Oct 16th 2008 01:22:41
The cleandrive script is the simplest and is run on the current OS in the /
(root) directory with root id.
dd if=/dev/zero of=/0bits bs=20M
rm /0bits
All it does is create a file of nulls until it files all the unsed space,
then deletes the file. Very simple, but no feedback, so it just sits there
until done.
The cleandrive5 and cleandrive6 are also run from the root of the OS. On
using the same dialog gui as the g4l and the same progress bar, the other
users xdialog gui. In both cases the dialog or xdialog needs to be
installed on the system. Additionally, the jetcat-mod program is also used
to produce the graph.
If it can do the image in 4 minutes with your hardware, I don't think it
would take long to clear the unused space. My machines have a 40GB linux
root partition, and the drives only get about 50Mb speed after cache is
used. So I wanted to monitor the progress.
There is also a blank6.exe program that is run on windows partitions to do
the same thing, but it has to create multiple 2GB files in case the
filesystems is FAT.
I once did a clean install of Fedora Core 3 onto an 80GB disk. Did an
image, and it was 12.5GB, then cleaned unused sectors, and the image went
down to about 2GB. But the used space was only about 6GB, so the 74GB of
random data in unused sectors took up about 10G of space.
I also use the cleandrive5 or cleandrive6 script in the /boot partition,
but it is so small probable doesn't make a difference. Those scripts create
the file in the directory you are in. Additionally, he jetcat-mod must be
the the / (root) directory, since it is hard coded into the scripts, but
they could be modified if you want to place the file elsewhere.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: v0.25 and up 515% slower
by Frank - Oct 17th 2008 01:46:03
Ok well did tests, all worked as expected. Cleandrive completed in 1mins43s
at 86megs/sec, then lzop backed up 14gigs in 3mins32sec and gzip in
7mins02sec (used to be 8mins37sec). The lzop file is 2.655gigs and the gzip
is 2.096gigs.
Interestingly, I still get my 26.7% bigger lzop file. But I see how it
works and how to clean efficiently the partition, I am happy you moved gzip
to bb so that I asked about it and then learned new stuff and things will
be even better now! :) tnx for the help!
[reply]
[top]
[»]
Re: v0.25 and up 515% slower
by msetzerii - Oct 17th 2008 10:58:44
Glad it worked out. Not sure why the gzip does so much better compression
with the data you have. I know that gzip does still do better compression
on already compressed files.
Additionally, perhaps gzip has code that makes better use of 64 bit cpus
or multi-cores.
If nothing else, G4L has been a lot of learning for me, and I've tried to
pass some of it onto my students.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
extract single file from g4l image ?
by joa - Sep 30th 2008 10:57:46
hi guys, first at all, thank you for this great software, i been using g4
since 2 years ago with no problems, im just wondering if there is a way to
extract a single file from a g4l image file ?
thx
[reply]
[top]
[»]
Re: extract single file from g4l image ?
by msetzerii - Sep 30th 2008 15:53:21
The first question is what kind of file system is the image?
I've done it once via the following method?
First, uncompress image file, and then mount the image file as a loop
device. This should work with most file systems, but would not work with
LVM partition with extra work. That is for partition backup, would have to
look into doing it with a disk image.
At the worst, one could always to a restore to a new disk, and then copy
the individual file, but that would require a disk of the same size or
larger.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
G4l issue
by zi247 - Sep 24th 2008 13:57:17
I am using g4l 0.26 and it seems everytime I try to do a restore it gets to
about 99.95% and I get a error that says. could not reply from control
Connection: Connection reset by peer. Any ideas?
[reply]
[top]
[»]
Re: G4l issue
by msetzerii - Sep 24th 2008 15:33:44
All the times I have heard of this issue, it turns out to be that the drive
is smaller than the original. The drives might appear to be exactly the
same, but you have to double check the exact sizes.
In my computer lab, I have 20 identical machines, but one of the drives is
about 25M smaller than the other drives.
They all have 255 heads, 63 sectors, but the one disk has 9726 cylinders
instead of the 9729 of the others.
I woudl suggest double checking what fdisk -l /dev/(device) reports for
the original disk and the new one.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
ftp://amd64gcc.dyndns.org/g4l-v0.27alpha20.iso
by msetzerii - Sep 10th 2008 02:02:12
ftp://amd64gcc.dyndns.org/g4l-v0.27alpha20.iso
Put a message out on adding ssh and encryption to sourceforge and
freshmeat hoping to get some input, but been over 2 weeks with nothing.
Here is the latest alpha I have worked with, and it seems to work just
fine, but I've got limited options, and I wrote the code, so it may not be
clear to others.
The ssh option is added to the local use menu, and uses the sshfs to make
the connection, and currently requires the inputting of the whole command.
It works with Fedora 6 and 8, but can not speak for other setups.
The encryption is only on the Network menu and is either on or off. The
gpg and passphrase file must be in the root directory.
I did create a very simple getkey.sh script to allow copying these files
from an USB device, but one could also just use ncftp of other methods.
It also has some newer kernels. 2.6.26.3 and 2.6.26.5, but those are not
the defaults. The .4 was only out for a short time before the .5 replaced
it, so I skipped keeping it.
The dev file is also in that same directory.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Adding sshfs and aes-pipe to g4l
by msetzerii - Aug 23rd 2008 11:26:09
Currently working on adding two more options to g4l.
1. Adding option to local image copy to mount a remote directory via
sshfs
/mnt/local, and then do backups via the ssh connection instead of using
ftp.
I have tested the process on one systems, and got a speed of about
20MB/sec
overall. A full disk image via ftp usually averages to 30MB/sec, but
that
includes the linux partition that is about 80% free space. So, there
might
be even less of a performance hit, but this adds an encrypted
connection
option, and also provides a method for those that can not use ftp. It
uses
the sshfs, which may or may not work with all servers. I've used it
on
Fedora 6 and Fedora 8 machines.
2. Adding an option on the Network menu to created encrypted image
files
using aespipe. This process is described below. It requires creation
of
a mykey1.gpg file and a passphrase file. Currently, the cd has a set
of
these files for testing. The ideal would be to create a script that
would
allow for getting these files onto the the ramdisk. This could be
download
from an ftp server or from flash device or perhaps a ssh connection.
At
the moment, this is just on the network option, and is a toggle.
I've currently made 14 versions of 0.27 so far.
At the moment, I've modified a copy of the regular g4l script g4l30b
to
include these options.
If others are interested in testing, and making comments on what is
the
best way to further implement the aespipe option.
Added option to use aespipe to created encrypted image files.
Currently only available on network option.
Encryption via aespipe.
AESPIPE by Jari Ruusu
http://sourceforge.net/projects/loop-aes/
This was recommended by a user, and seems to work fine.
Have had some communication with the author.
The CD image has a test set of key and passphrase file created as a demo,
but should not be used for real encryption, since they are publicly
available.
/mykey1.gpg and /passphrase
The method used to create these came from the aespipe README file.
== section of README
Create 65 random encryption keys and encrypt those keys using gpg.
Reading
from /dev/random may take indefinitely long if kernel's random entropy
pool
is empty. If that happens, do some other work on some other console
(use
keyboard, mouse and disks). Use of gpg-encrypted key file depends on
encrypted swap.
head -c 2925 /dev/random | uuencode -m - | head -n 66 | tail -n 65
\
| gpg --symmetric -a >mykey1.gpg
=== emd
The program using the aespipe to encrypt and decrypt using the
following
additional code.
AESPIPE=" aespipe -K /mykey1.gpg -P /passphrase -G / | "
AESPIPEd="| aespipe -K /mykey1.gpg -P /passphrase -G / -d "
These variables are set if encryption is toggled on, and inserted
into
the process. Otherwise the values are blank and have no effect on the
process.Currently the names of these files are hard coded, so one would
need to create you own key file and passphrase, and replace these.
You files could be downloaded from a secure location after booting
from
the cd by various options and others.
ncftp -u user -p password ipaddress
get mykey1.gpg
get passphrase
quit
That puts the files on an ftp site, which you may or may not want to
do.
Insert an USB device
mount /dev/sdx /mnt/local
cp /mnt/local/mykey1.gpg
cp /mnt/local/passphrase
umount /mnt/local
At this point, I will assume that users wanting to use encryption
will
be more than capable of creating these files, and setting them up.
Later this could be further expanded and automated more.
Recommend testing you mykey1.gpg and passphrase files before making
an
image file, since if it doesn't work the image will not.
Following script is a quick test. It just creates a file with random
data, encrypts it, and then decrypts it. If the original file and
decrypted
file produce the same results, it worked.
#!/bin/sh
#
#
## Michael Setzer II - Program Modifications
## Set tab = 2 spaces for formatting
dd if=/dev/urandom of=./aes1.test bs=512 count=40
aespipe -K ./mykey1.gpg -P ./passphrase -G ./ ./aes2.test
aespipe -K ./mykey1.gpg -P ./passphrase -G ./ -d ./aes3.test
cmp -l ./aes1.test ./aes3.test
Very limited testing has been done so far.
With a 20GB NTFS partition.
None Encrypted image took 15m 55s
Encrypted image took 16m 35s
That works to be just about a 4% increase, but would depend on
hardware,
so your results may be higher or lower. Test done with P4 2.8
machine.
Tis is only added to my working 0.27 alphas, and I'm currently a level
14
Have not yet released any of these publicly.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Cannot see partitions
by molinos - Aug 18th 2008 19:42:54
When I boot from the G4L cd I get the following messages at the start of
the boot process. Once the G4L shell comes up, I do not see any partition
information. I am trying to backup a Dell Poweredge 2850 with Ultra 320
SCSI 146GB 15K drives. It has a LSI Logic controller.
GDT-HA: Storage RAID Controller Driver. Version: 3.05
GDT-HA: Found 0 PCI Storage RAID Controllers
3ware Storage Controller device driver for Linux v1.26.02.002.
3ware 9000 Storage Controller device driver for Linux v2.26.02.010.
nsp32: loading...
ipr: IBM Power RAID SCSI Device Driver version: 2.4.1 (April 24,
2007)
RocketRAID 3xxx/4xxx Controller driver v1.3 (071203)
stex: Promise SuperTrak EX Driver version: 3.6.0000.1
Driver 'sd' needs updating - please use bus_type methods
Driver 'sr' needs updating - please use bus_type methods
SCSI Media Changer driver v0.25
Driver 'ch' needs updating - please use bus_type methods
I2O subsystem v1.325
i2o: max drivers = 8
I2O Configuration OSM v1.323
I2O Bus Adapter OSM v1.317
I2O Block Device OSM v1.325
I2O SCSI Peripheral OSM v1.316
I2O ProcFS OSM v1.316
Fusion MPT base driver 3.04.06
Copyright (c) 1999-2007 LSI Corporation
Fusion MPT SPI Host driver 3.04.06
mptbase: ioc0: Initiating bringup
mptbase: ioc0: WARNING - Unexpected doorbell active!
mptbase: ioc0: ERROR - Doorbell ACK timeout (count=4999),
IntStatus=80000001!
mptbase: ioc0: ERROR - Diagnostic reset FAILED! (102h)
mptbase: ioc0: WARNING - NOT READY!
mptbase: ioc0: ERROR - didn't initialize properly! (-1)
mptspi: probe of 0000:02:05.0 failed with error -1
mptbase: ioc1: Initiating bringup
mptbase: ioc1: ERROR - Doorbell ACK timeout (count=4999),
IntStatus=80000000!
mptbase: ioc1: ERROR - Diagnostic reset FAILED! (102h)
mptbase: ioc1: WARNING - NOT READY!
mptbase: ioc1: ERROR - didn't initialize properly! (-1)
mptspi: probe of 0000:02:05.1 failed with error -1
--
Molinos
[reply]
[top]
[»]
Re: Cannot see partitions
by msetzerii - Aug 18th 2008 22:04:57
What does the /proc/partitions file show??
What devices do you see in the /dev directory??
There will be a number of devices for other devices, but does it show any
disk devices??
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Cannot see partitions
by msetzerii - Aug 18th 2008 22:36:23
I further checked the .config for the kernel build, and all the options for
the I2O and LSI are set on, so unless the option falls under something else
in the kernel build, it seems to be set on.
I have had some other raid controllers that required special nods to be
created. The ida, ataraid, and cciss required special subdirectories, and
nods for support. I did find a nod that is required for some LSI
controllers. The latest g4l used busybox mdev to create nods, and it works
find for all the devices I have access to. I don't have any of the systems
with ida, ataraid, or cciss controllers, so I have left the manually
created nods for those until I can confirm the mdev can also create those.
I did find the following on one web page about LSI nods.
mknod /dev/mptctl c 10 220
So, if that doesn't get created it might need to be added to the g4l
build. It also shows /proc/mpt directory.
Also, exactly which version of g4l are you using, and what does the
/proc/partitions and /dev contain on the used OS show?
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Cannot see partitions
by molinos - Aug 19th 2008 12:13:24
Here is my /proc/partitions
major minor #blocks name
1 0 65536 ram0
1 1 65536 ram1
1 2 65536 ram2
1 3 65536 ram3
1 4 65536 ram4
1 5 65536 ram5
1 6 65536 ram6
1 7 65536 ram7
1 8 65536 ram8
1 9 65536 ram9
1 10 65536 ram10
1 11 65536 ram11
1 12 65536 ram12
1 13 65536 ram13
1 14 65536 ram14
1 15 65536 ram15
Also my /dev has one entry for hda and md0, but nothing else that looks
like a drive.
-bash-3.2# ls /dev |grep -v loop |grep -v tty |grep -v usb |grep -v vc
ataraid flush mice nbd2 ptmx ram5
cciss full microcode nbd3 pts ram6
console fuse mouse0 nbd4 ram0 ram7
cpu0 hda mptctl nbd5 ram1 ram8
cpu_dma_latency hpet msr0 nbd6 ram10 ram9
dac960_gam i2octl nbd0 nbd7 ram11 random
device-mapper ida nbd1 nbd8 ram12 revalidate
discover interfaces nbd10 nbd9 ram13 root
err kmem nbd11 network_latency ram14 rtc
event0 kmsg nbd12 network_throughput ram15 uinput
event1 md0 nbd13 null ram2 urandom
event2 megadev0 nbd14 nvram ram3 zero
fd0 mem nbd15 port ram4
Here is my /proc/mpt/
-bash-3.2# ls /proc/mpt
summary version
-bash-3.2# cat /proc/mpt/summary
-bash-3.2# cat /proc/mpt/version
mptlinux-3.04.06
Fusion MPT base driver
Fusion MPT SPI host driver
Fusion MPT FC host driver
Fusion MPT SAS host driver
Fusion MPT ioctl driver
Fusion MPT LAN driver
-bash-3.2#
I am using version 0.26a.
I am going to try to include a screen shot of my sys info screen, but it
is basically blank.
I will look into your other comments and get back to you soon.
Thanks for your quick response.
--
Molinos
[reply]
[top]
[»]
Re: Cannot see partitions
by molinos - Aug 19th 2008 12:24:58
WOW did that ever post bad...
Should have looked like this:
> Here is my /proc/partitions
> major minor #blocks name
>
> 1 0 65536 ram0
> 1 1 65536 ram1
> 1 2 65536 ram2
> 1 3 65536 ram3
> 1 4 65536 ram4
> 1 5 65536 ram5
> 1 6 65536 ram6
> 1 7 65536 ram7
> 1 8 65536 ram8
> 1 9 65536 ram9
> 1 10 65536 ram10
> 1 11 65536 ram11
> 1 12 65536 ram12
> 1 13 65536 ram13
> 1 14 65536 ram14
> 1 15 65536 ram15
>
>
> Also my /dev has one entry for hda and
> md0, but nothing else that looks like a
> drive.
>
-bash-3.2# ls /dev |grep -v loop |grep -v tty |grep -v usb |grep -v vc
> ataraid
> cciss
> console
> cpu0
> cpu_dma_latency
> dac960_gam
> device-mapper
> discover
> err
> event0
> event1
> event2
> fd0
> flush
> full
> fuse
> hda
> hpet
> i2octl
> ida
> interfaces
> kmem
> kmsg
> md0
> megadev0
> mem
> mice
> microcode
> mouse0
> mptctl
> msr0
> nbd0
> nbd1
> nbd10
> nbd11
> nbd12
> nbd13
> nbd14
> nbd15
> nbd2
> nbd3
> nbd4
> nbd5
> nbd6
> nbd7
> nbd8
> nbd9
> network_latency
> network_throughput
> null
> nvram
> port
> ptmx
> pts
> ram0
> ram1
> ram10
> ram11
> ram12
> ram13
> ram14
> ram15
> ram2
> ram3
> ram4
> ram5
> ram6
> ram7
> ram8
> ram9
> random
> revalidate
> root
> rtc
> uinput
> urandom
> zero
>
> Here is my /proc/mpt/
>
> -bash-3.2# ls /proc/mpt
> summary version
> -bash-3.2# cat /proc/mpt/summary
> -bash-3.2# cat /proc/mpt/version
> mptlinux-3.04.06
> Fusion MPT base driver
> Fusion MPT SPI host driver
> Fusion MPT FC host driver
> Fusion MPT SAS host driver
> Fusion MPT ioctl driver
> Fusion MPT LAN driver
> -bash-3.2#
>
> I am using version 0.26a.
>
>
> I am going to try to include a screen
> shot of my sys info screen, but it is
> basically blank.
>
> I will look into your other comments and
> get back to you soon.
>
> Thanks for your quick response.
>
--
Molinos
[reply]
[top]
[»]
Re: Cannot see partitions
by molinos - Aug 19th 2008 14:34:53
Also, my MPT bios is 5.06.06
--
Molinos
[reply]
[top]
[»]
Re: Cannot see partitions
by msetzerii - Aug 19th 2008 20:29:45
What OS are you running on the machine?
If Linux or other Unix like OS, what does it show in the /proc/partitions
file, and what does df show.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Cannot see partitions
by msetzerii - Aug 19th 2008 21:42:42
ftp://amd64gcc.dyndns.org/g4l-v0.27alpha13.iso
Is the latest working copy I have, and just added that /dev entry that I
found.
So, it might show something additional.
Until we can get something to show up in the /proc/partitions, we have a
problem.
You haven't said what OS you are running on the system, so getting info on
how it sees the drive. If not, a Linux system, it might be interesting to
boot from a full live cd system like knoppix, and see if it sees the drive,
and what it has in the /proc/partitions and /dev.
Also, in posting, using html and using the br instead of the p reduces
blank lines.
The 0.27 is working on adding aespipe to create encrypted images, and also
has sshfs, to allow for using an ssh connection instead of ftp, but the
process is manual at the moment. So they both can work, but require
some extra steps before the imaging.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Cannot see partitions
by molinos - Aug 20th 2008 12:36:17
I am running Trustix (uname -a : Linux host1 2.6.21.5-4q1smp #4 SMP Mon Aug
13 18:33:22 ADT 2007 i686 i686 i386 GNU/Linux)
on the box I am trying to backup.
Here is my /proc/partitions
cat /proc/partitions
major minor #blocks name
9 0 573487872 md0
8 0 71687325 sda
8 1 104391 sda1
8 2 48532365 sda2
8 3 10498477 sda3
8 4 1 sda4
8 5 10442218 sda5
8 6 2104483 sda6
8 16 143374740 sdb
8 17 143372061 sdb1
8 32 143374740 sdc
8 33 143372061 sdc1
8 48 143374740 sdd
8 49 143372061 sdd1
8 64 143374740 sde
8 65 143372061 sde1
8 80 143374740 sdf
8 81 143372061 sdf1
There are over 500 entries in my /dev for hd*
I am continuing to look into it from my end also. Again thanks for all
your help.
--
Molinos
[reply]
[top]
[»]
Re: Cannot see partitions
by molinos - Aug 20th 2008 14:17:05
df -k
Filesystem 1K-blocks . . . . Used . . . Available . Use% Mounted on
/dev/sda2 .. 48530864 . 4696852 .. 43834012 . 10% . /
/dev/sda1 ..... 101086 .... 51144 ....... 44723 . 54% . /boot
/dev/sda3 .. 10498104 .... 83384 .. 10414720 ... 1% . /var/log
/dev/md0 . 573470364 .. 692396 . 572777968 ... 1% . /store
/dev/sda5 .. 10441852 .. 345872 .. 10095980 ... 4% . /store/tmp
--
Molinos
[reply]
[top]
[»]
Re: Cannot see partitions
by taternuts - Sep 19th 2008 15:20:04
I know this may be a little late, but....
I've done about a half dozen Power Edge 2850's with G4L. I also had a
problem with the LSI/PERC RAID. But it worked with G4L version 0.23, using
the kernel "A: bzImage6 Previous...".
If it helps any, this same option did not work using G4L 0.24 or 0.26
until I updated the System Bios and the RAID controller firmware. Once I
did that, both the newer versions of G4L work fine with the 2850.
Unfortunately I updated both the BIOS and firmware before testing again,
so I'm not sure which one actually helped. But we still have a few 2850's
in the warehouse, so the next time I have to deploy one I'll test a bit
more.
So, you may want to update the BIOS to ver. A07, dated 2008/5/23 and the
RAID firmware to 5A2D,A18 dated 2008/10/2007. Or try ver. 0.23 of G4l,
using the A: kernel.
HTH
--
- tater -
[reply]
[top]
[»]
Nvidia RAID
by Laserbait - Aug 8th 2008 17:38:19
Hi there!
I was wondering if NVraid was supported natively yet? I have a Nvidia
Nforce4 RAID 0+1 (4x500GB Seagate drives) and I'd like to image the OS
partition. In past versions, it only saw the drives as individual drives,
not as a RAID set.
Thanks!
--Joseph
--
SMP-aholic
[reply]
[top]
[»]
Re: Nvidia RAID
by Laserbait - Aug 8th 2008 17:42:47
> Hi there!
>
> I was wondering if NVraid was supported
> natively yet? I have a Nvidia Nforce4
> RAID 0+1 (4x500GB Seagate drives) and
> I'd like to image the OS partition. In
> past versions, it only saw the drives as
> individual drives, not as a RAID set.
>
> Thanks!
>
> --Joseph
Also, I forgot to ask on moire question. Is there is more information on
using multicast to improve restore times? The docs don't say anything
about it, probably because they are from v.16. :)
--
SMP-aholic
[reply]
[top]
[»]
Re: Nvidia RAID
by msetzerii - Aug 8th 2008 19:13:43
Question 1:
What is the contents of cat /proc/partitions?
What is the output of fdisk -l
It may show the individual drives in there raw mode hdx or sdx, or it may
show the raid device as an mdx.
If it shows the mdx, it might be that you could use this to backup the
raid, but I would want to do a test to make sure it works. Regardless, you
should be able to boot the individual drives as well?
Example: I got some old HP servers from a Hotel, and they had scsi setup
with multiple 9GB removable drives. I took one system, and set it up with 2
of the 9GB drives, and installed and updated linux on them. I then made
images of both drives, and restored them to two other sets of drives one by
one, and they worked fine in the other machines.
On the question of Multi-cast. The g4l cd includes udpcast, and I use it
all the time. I generally make g4l backups of the image to the ftp server,
but when I want to update all other machines in the lab, udpcast does it at
one time. It can use the image files created by g4l or it can directly
image the disk. You can make cd image at the udpcast web site to support
your systems. The udp-sender udp-receiver programs are on the cd, but only
the udp-receiver program is included as part of the menu. The udp-sender
would be used on the ftp server with the image, or the machine being used
to copy from.
What I have actually done in my lab, is add the g4l, udpsender, and
udprecieiver images to my grub menu, so I can boot directly to them. Used
the udpcast option to build custom cd images for my hardware, and then just
copied the necessary files to the /boot directory.
Sometimes newer hardware isn't supported by the udpcast cd image
generator, but it supports a lot, and is updated from time to time. If g4l
support hardware the udpcast doesn't, one could run the commands directly,
so it can be done, but more detail would need to be worked out.
So, at the least, you should be able to back the individual drives that
make up the raid system if not the raid itself, and multi-cast can be done
via the udpcast program either as part of g4l or seperate.
Finally, is your Raid system raid 0 or raid 1?
If it is Raid 0, you would need to backup both drives, since the data is
spanned or stripped across both. If it is raid 1 mirror drives, you could
probable backup just the primary drive, since the secondary is just a copy.
If a higher level raid, you would probable need to back up drives.
Technically, you could backup all but one drives, and count on the raid to
rebuild it, but I wouldn't trust that kind of backup.
Hope I answered you questions at least to a degree. You might want to
further clarify the specifics, and provide the output of how your system
sees the drives. Additionally, you might want to look at the contents of
the /dev directory, and see what disk devices it list. /dev/sdx, /dev/hdx,
and possible /dev/mdx. The x of course being various letters /dev/sda,
/dev/sda1, etc.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: Nvidia RAID
by Laserbait - Aug 9th 2008 00:01:40
Hey Mike,
Thanks for the info. I'll check out the /dev items when I get home. The
Nvidia RAID is a RAID 0+1 (sometimes called 10), in which there are 4
disks. The drives are split into 2 pair. The drives in each pair are
striped together, and then the 2 pair are mirrored together. You get most
of the speed of a RAID0 but with redundancy.
--
SMP-aholic
[reply]
[top]
[»]
g4l on multi-block (RAIDed) devices
by thorae - Aug 1st 2008 18:43:55
Hi, I'm attempting to create and restore ghost images of a system with 2 or
more block devices. I'm currently in the need to install a linux based
software system multiple times. There is a base release and subsequent
releases are installed as upgrades. So successfully storing a base image on
an FTP server and restoring as needed on different boxes will save me a lot
of time.
Here's the twist: the software makes use of a lot of disk space so each
box consist of 2 or more large disk drives. All partitions (including /boot
and /) are mounted on the first disk device and the rest are set up on a
RAID partition. Now how do I use g4l to create and restore a ghost image of
this? Do the individual disk devices need to be backed up separately?
Because it only allows me to select one device at a time.
On boot up after restoring I get:
"An error occurred during the file system check.
Dropping you to a shell....
fsck.ext3: Unable to resolve 'Label=/store' "
Lastly, it's not because the backup drives is smaller than the restore
drives because I even tried restoring on the same system and still got the
above error.
Any ideas anyone?
[reply]
[top]
[»]
Re: g4l on multi-block (RAIDed) devices
by msetzerii - Aug 1st 2008 20:37:34
First, you didn't mention what level of Raid? are we talking spanning or
mirror or higher level?
The label information is in the partition table from my understanding.
I've discovered that some OS's are using UUID in boot file (grub.conf or
menu.lst) and if restored to different physical disks, this needs to be
reset for the boot process to match up.
On the issue of multiple drives. What options are you see? My college has
some older HP servers donated that had multiple 9GB drive. I setup one
system using 4 spanned drives, and then made images of each drive. Then
restored the images to each of the 4 drives in the other 2 machines, and it
worked fine.
So, it might be that you need to do images of each drive, and restore
them. Also, what is the OS you are using? What does the grub.conf or
menu.lst show? What do you find in /proc/partitions after booting from the
g4l cd.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
g4l can't recognise Ultrabay 2000 usb HDD
by mr4 - Jul 23rd 2008 10:15:31
First of all, top TOP piece of software!
I have been using it for more than a year and it saved my skin on a
numerous occasions.
One suggestion may be is if I could secure my backups with a password so
to prevent unauthorised access to them. One thing missing was the fsck
tools, which I am glad are included in the new (v0.26) version - I have
been using v0.23 up until now.
About my particular problem - I normally do ftp-based backups (on a
separate machine running fc8), but recently acquired 160GiB HDD which would
allow me to (hopefully) do local backups - and here is where I have a
little difficulty.
The HDD in question is 2.5" ATA-6 drive, which works without a hitch
when I plug it into my thinkpad (A31) using the 2nd HDD bay slot.
Since I want to make a 'local' backup to all my PCs, not just the
thinkpad, I use the Ultrabay 2000 which has usb connector allowing me to
connect the disk to all my desktop PCs (it also has PCMCIA card as well but
that won't work on Desktop PCs).
The problem is that as soon as I connect the usb port to the computer I
get the message that Mass Storage Device is recognised, but it stops there,
so g4l ultimately fails to recognise the hard disk plugged in through the
usb port via the Ultrabay 2000 device. I am using g4l v0.23 - just
downloaded the newest (v0.26) version, but not sure if it is going to
work.
I have even tried doing the same thing within a 'proper' Linux system -
fc8 in this case - but get no luck there either: same message - it
recognises a 'Mass Storage USB device' and then does not map it to a proper
/dev/ device which I can use. Any help with this would be greatly
appreciated!
I know a work-around of sorts - to discard the Ultrabay 2000 altogether
and plug the drive directly into the IDE/ATA interface on the motherboard
of the desktop PC, but that is very inconvenient since I have to do that
for every Desktop PC to which I need a backup done!
--
mr-4
[reply]
[top]
[»]
Re: g4l can't recognise Ultrabay 2000 usb HDD
by msetzerii - Jul 23rd 2008 13:18:20
A couple of suggestions.
Try the 0.26, it has the busybox mdev setup.
Then do the following:
1: Press Enter until you get to regular command line.
2: Check /proc/partitions
cat /proc/partitions
3: Plug in the device and see what happens.
4: Recheck /proc/partitions
Note: The g4l script reads the /proc/partitions file at the beginning, so
the device needs to be plugged in early, or the g4l script needs to be
restarted.
5: If it isn't recognized, check out dmesg
dmesg | more
6: Try to run mdev -s manually:
mdev -s
7: Try fdisk -l and see if it reports anything
The only other thing I can think of, is to try plugging in the device
before booting the machine. I had one usb device that would only work this
way, don't know why.
Thanks for the comments.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: g4l can't recognise Ultrabay 2000 usb HDD
by mr4 - Jul 23rd 2008 23:58:36
OK, here is what I did:
1. and 2. - apart from ramX partitions and my proper HDD (attached to the
IDE/ATA on the motherboard) nothing shows up.
3. This is what I get:
==start==
hub 1-0:1.0: unable to enumerate USB device on port 1
usb 2-1: new full speed USB device using ohci_hcd and address 3
usb 2-1: configuration #1 chosen from 1 choice
scsi3 : SCSI emulation for USB Mass Storage devices
usb-storage: device found at 3
usb-storage: waiting for device to settle before scanning
usb 2-1: reset full speed USB device using ohci_hcd and address 3
usb 2-1: reset full speed USB device using ohci_hcd and address 3
usb 2-1: reset full speed USB device using ohci_hcd and address 3
usb 2-1: reset full speed USB device using ohci_hcd and address 3
usb-storage: device scan complete
==end==
As you can see there are two errors: 1 at the very beginning - "unable to
enumerate USB device on port 1" and 1 (repeated 4 times) at the end -
"reset full speed USB device using ohci_hcd and address 3". Don't know what
any of those mean.
Here is what I normally get (in terms of usb output) when I plug usb disk
(memory stick in this case), which is recognised by the system:
==start==
usb 1-4: new high speed USB device using ehci_hcd and address 5
usb 1-4: configuration #1 chosen from 1 choice
scsi4 : SCSI emulation for USB Mass Storage devices
usb-storage: device found at 5
usb-storage: waiting for device to settle before scanning
scsi 4:0:0:0: Direct-Access TREK TDSWIPE G6 2.00 PQ: 0 ANSI:
0 CCS
sd 4:0:0:0: [sda] 2003904 512-byte hardware sectors (1026 MB)
sd 4:0:0:0: [sda] Write Protect is off
sd 4:0:0:0: [sda] Mode Sense: 0b 00 00 00
sd 4:0:0:0: [sda] Assuming drive cache: write through
sd 4:0:0:0: [sda] 2003904 512-byte hardware sectors (1026 MB)
sd 4:0:0:0: [sda] Write Protect is off
sd 4:0:0:0: [sda] Mode Sense: 0b 00 00 00
sd 4:0:0:0: [sda] Assuming drive cache: write through
sda: sda1
sd 4:0:0:0: [sda] Attached SCSI removable disk
sd 4:0:0:0: Attached scsi generic sg0 type 0
usb-storage: device scan complete
usb-stor-scan used greatest stack depth: 1852 bytes left
==end==
After that the device is mapped to /dev/sda with no problems...
6. when I run it nothing happens and /proc/partitions doesn't show any
difference (no errors from 'mdev -s' either).
7. Same - no difference
When I try to boot up with the usb drive attached I see no difference
(dmesg are the same, but this time at the beginning).
On another note, with regards to the comment I posted earlier (below) this
is what I am also getting in dmesg (don't know if it is an error or not):
Driver 'sd' needs updating - please use bus_type methods
Driver 'sr' needs updating - please use bus_type methods
Driver 'ch' needs updating - please use bus_type methods
--
mr-4
[reply]
[top]
[»]
Re: g4l can't recognise Ultrabay 2000 usb HDD
by msetzerii - Jul 23rd 2008 21:38:24
Followup question on this device.
Is it a special drive for the Lenova laptops?
I recently got a 100GB USB 2 1/2 drive off ebay for about $50 for my mom's
machine, so she can do backups to it. It works fine, and is recognized.
The kernels used by g4l are built from the kernel.org source, and I
checked the .config and looked at lines with IBM and ULTRA, but didn't see
anything that stood out. So, it might be that this is a device that just
works with the Lenova or requires a driver that is not included with the
kernel.org source.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: g4l can't recognise Ultrabay 2000 usb HDD
by mr4 - Jul 23rd 2008 22:57:11
> Followup question on this device.
> Is it a special drive for the Lenova
> laptops?
No, it is not as far as I am aware - it is described here:-
http://www-307.ibm.com/pc/support/site.wss/document.do?lndocid=MIGR-4JSSC8
and here:- http://www.thinkwiki.org/wiki/UltraBay
When I plug it in in a 'normal' (i.e. Desktop) PC using Windows XP - it is
recognised and the drive is mapped as usual.
> I recently got a 100GB USB 2 1/2 drive
> off ebay for about $50 for my mom's
> machine, so she can do backups to it. It
> works fine, and is recognized.
> The kernels used by g4l are built from
> the kernel.org source, and I checked the
> .config and looked at lines with IBM and
> ULTRA, but didn't see anything that
> stood out. So, it might be that this is
> a device that just works with the Lenova
> or requires a driver that is not
> included with the kernel.org source.
I don't know what is 'kernel.org' source, but I read on various sources
about the need to load the 'ide-scsi' device (this is also suggested in the
2nd link I posted above), but that is applicable only to 2.4 kernels, not
the 2.6 version as this is already 'built-in'. Another suggestion I tried
is to include 'hdc=ide-scsi' in the kernel bootup options. I did try that
(including g4l menu by pressing the tab key and adding 'hdc=ide-scsi' - no
good, nothing happened). I also tried 'fdisk -l', but it lists just the
partitions on my cd-rom. Will try your other suggestions tomorrow and will
reply to it separately.
Many thanks for your help - as I said this is a great piece of software -
I looked at various different (and ultimately much more expensive)
'solutions' but nothing stood out more than g4l in terms of flexibility,
ease of use and reliability. As I pointed out in my previous comment - the
one thing which is probably missing is locking backups with a password to
prevent unauthorised access - don't know if that is easy/possible to
implement in future versions.
--
mr-4
[reply]
[top]
[»]
Re: g4l can't recognise Ultrabay 2000 usb HDD
by msetzerii - Jul 24th 2008 01:44:16
The link that you mention shows it as being a special drive plug in bay
from the link I saw. Are taking the drive out of the bay, or how are you
connecting it to a regular desktop?
http://kernel.org
Is the site that contains the full public source code for the linux
kernels. You can download the various kernel source code, and then build a
kernel from that code base. I try to build into my kernels all the
available network and harddrive support that isn't listed as experimental
or dangerous.
I know that I have used regular hard drives and laptop hard drives in USB
cases with g4l, so know sure what is different about this drive. While have
to look at it a little more.
As far as the question on security. I have kind of left it up to the
userid and password for the security issue. The original g4l only worked
with anonymous userid with no password, and I didn't like that, so I added
the ability to use any userid on a system and password. I also use an id
for creating backs on my systems, and let others know the id so they can
restore images. I generally, change the owner of images to root afterwards,
so they can download and restore, but not overwrite the images by accident.
I would have to run a lot of test, and find an encryption program that
could work on the fly, and have little reduction in the speed. Don't know
enough on encryption options. Wouldn't due to have something that would
encrypt during the process, but then not be able to decrypt. Also, if it
adds to much to the time.
I don't see a problem in the transfer, since one would have to capture the
entire image to be able to get the data.
If I find more info, or if you do, drop a note. I've seen similar plug in
devices on dell notebooks, but as it says on that web page, it is handled
by the BIOS, so the OS doesn't do anything. But if you connect it to a
regular desktop, how does that the the information in the BIOS?
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: g4l can't recognise Ultrabay 2000 usb HDD
by mr4 - Jul 24th 2008 09:21:00
I think 'special' is the wrong word - the device is made by IBM and its
main purpose is to add extra devices (CD-RW, DVD-R, DVD-RW, FDD, HDD etc)
to (mostly) Thinkpad laptops, which do not have it in their configuration.
It is completely universal that it accepts a standard IDE/ATAPI/ATA-capable
devices and connects them to a computer.
Ultrabay 2000 has two different connectors - usb (can be plugged in to
desktops as well as notebook computers having USB 1.1 and USB2.0 ports) and
PCMCIA (for notebooks only as desktops can't use those to my knowledge)
ports to connect itself to a PC.
As I said previously I was able to successfully connect the device to a
desktop PC running on Windows XP using the usb port - no problems there.
The problems arise only on linux-type system, which makes me think that the
drivers supporting this kind of device are not there.
As far as security on g4l backups goes - are you saying that in 0.26 I can
use userid:passwords to store backups? I thought these are when I connect
to ftp server? Is this something new in 0.26 or have I missed something?
If you are interested in encryption-on-the-fly one VERY good source is
loop-aes (loop-aes.sourceforge.net or
http://koti.tnnet.fi/jari.ruusu/linux/) - it makes it possible to create
encryption drives (loop devices) and whatever is recorded/written there is
encrypted on-the-fly with very little overhead in real time. I am using 2
partitions like this and haven't had a single problem for 4 years. It works
only on Linux though.
--
mr-4
[reply]
[top]
[»]
Re: g4l can't recognise Ultrabay 2000 usb HDD
by msetzerii - Jul 24th 2008 13:53:32
In looking at the
http://www-307.ibm.com/pc/support/site.wss/document.do?lndocid=MIGR-4JSSC8
page, Under the hardware it only shows Thinkpad and NetVista
Under Supported Operating Systems
* Microsoft Windows 9
* Microsoft Windows 98
* Microsoft Windows NT 4.0
* Microsoft Windows 2000 for PCMCIA support
* Microsoft Windows 98 or Windows 2000 for USB support
It doesn't list XP or linux, but as you mentioned it does work with XP. If
you go to drive management, what kind of disk (physical) does it list?
As for the userid:password security, I am referring to the network backup
via FTP. This is used to make the connection via the ftp server, and thus
the file access would require this user id and password to access the file
via ftp. The OS might allow access to others, but that is a permission
issue.
I'll take a look at that program you recommended. Have to see what options
it needs passed, and how that works running from a cd image.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: g4l can't recognise Ultrabay 2000 usb HDD
by mr4 - Jul 24th 2008 18:56:00
> It doesn't list XP or linux, but as you
> mentioned it does work with XP. If you
> go to drive management, what kind of
> disk (physical) does it list?
'Portable' (the same type as Memory sticks). I am almost at the point of
resigning on this now. Will make a last try to see if it works via the
PCMCIA card at the weekend (will test this on the laptop first) and if it
does work will get PCMCIA (cardBus) -> IDE/ATA adapter and that is how I
will use it.
> I'll take a look at that program you
> recommended. Have to see what options it
> needs passed, and how that works running
> from a cd image.
The part which will be of most interest to you is aespipe (loop-aes
requires support in the kernel and additional code to be included in before
recompiling this kernel). aespipe is AES encrypting or decrypting pipe. It
reads from standard input and writes to standard output. It can be used to
create and restore encrypted tar or cpio archives. It can be used to
encrypt and decrypt loop-AES compatible encrypted disk images (the part I
use it most for).
It does not require additional modules and when compiled in produces
just two files (executable + sh script - all placed in /usr/bin) and a man
page.
Examples of usage:
tar cvf archive.aes --use-compress-program=bz2aespipe files... (for
compressing); and
tar xvpf archive.aes --use-compress-program=bz2aespipe (for
decompressing)
Also,
tar cvf - files... | bzip2 | aespipe -w 10 -K mykey1.gpg > archive.aes
(for compressing, using gpg key created/generated in advance); and
aespipe -d -K mykey1.gpg < archive.aes | bzip2 -d -q | tar xvpf - (for
decompressing).
As you can see, if you redirect the output pipe coming out of lzop
(while doing a backup) to the aespipe then you can create an encrypted file
on the fly so to speak. The opposite is also true when doing a restore.
For further info check
http://koti.tnnet.fi/jari.ruusu/linux/aespipe-v2.3d.tar.bz2 and check the
'README' file inside this archive!
--
mr-4
[reply]
[top]
[»]
Re: g4l can't recognise Ultrabay 2000 usb HDD
by msetzerii - Jul 25th 2008 03:32:04
To double check. Did you try hooking up the device before starting the
system. I had one USB external drive that would only work if it was
connected and powered on before the systems was booted up. The kernel has
all the modules build in, and activates at boot time. It may not be seeing
the device at boot and not loading the necessary device.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: g4l can't recognise Ultrabay 2000 usb HDD
by mr4 - Jul 25th 2008 09:33:16
I did - it didn't work. I have tried the PCMCIA interface yesterday and it
worked - no problem at all. As soon as I put the PCMCIA card in, the drive
was instantly recognised by g4l and assigned a device - 'sda' in this case.
Did 'cat /proc/partitions' and it was all there. Did 'mount' to see the
contents of the partition - no problem. So, if everything else fails I
might go for this.
Another option I am considering is changing the Ultrabay with something
else - ordered a 2.5" (P)ATA/SATA to USB2.0/eSATA caddy yesterday and will
test it when it arrives in a few days.
On a separate note, I have looked at your g4l30 script yesterday as I was
intrigued about the possibility of using aespipe to encrypt backups on the
fly.
I was surprised how flexible and configurable the g4l menu system is. It
should not be a problem to add an extra menu option for
encryption/decryption as well as specifying encryption key and then
redirecting the input/output to aespipe before saving. You are already
doing it, so adding an extra pipe in the middle between the ncftp/dd
input/output and the compression program (gz, lzop or tar) won't be a
problem.
If I manage to fix my Ultrabay woes I will do some testing on that front
during the weekend. Integrating aespipe into 4gl won't be a problem I don't
think - 'make' produces 1 executable (aespipe), one script (bz2aespipe) and
one man page, which could easily be copied to the appropriate directories
before the g4l iso file is made (I might also add basic gpg routines for
generating the gpg key if everything works).
--
mr-4
[reply]
[top]
[»]
Re: g4l can't recognise Ultrabay 2000 usb HDD
by msetzerii - Jul 25th 2008 10:50:18
Thanks for the update.
I started looking at the encryption also. I downloaded it, and built it
with no problem, but the instructions to build the key didn't work as
stated. First, uuencode wasn't installed on my system, had to search yumex
to find it was in a util package. Then it still didn't work. Had to switch
from /dev/random to /dev/urandom. Not sure what the diff is, but that
seemed to work. Did get into testing it yet.
To make the additiona simple at this stage, I was looking at just using
the special options to set a variable string that would include a pipe
symbol at the end. This way, it could be set to "" if not used, and would
do nothing, otherwise it could be set to the command followed with the
piping symbol to pass on the information.
Probable just add it to the lzop option for a start. Then do image with
and without it, and check the times. Then do a restore to make sure the
process works both ways.
I head back to my home on Monday, so I'll have direct access to my systems
instead of working from 7 timezones away.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: /random and /urandom devices
by mr4 - Jul 25th 2008 13:49:20
There is a massive difference between /dev/random and /dev/urandom -
/dev/random guarantees you good quality entropy, /dev/urandom does not.
With using /dev/random it takes a while (it blocks until you get the bits
you've requested, so don't panic, just be patient), it is not fast as it
needs to generate enough entropy bits. /dev/urandom, on the other hand,
does not guarantee you good quality entropy and will give you "anything
what's available" which is potentially dangerous as it could compromise the
quality of the generated key.
There are a few things you need to be wary of though - it would be wise to
provide an option to present the key file (I have mine generated in advance
and stored separately on a memory stick - for improved security), hence my
suggestion of including an option to specify the key file and not
generating it automatically.
Also, if you generate the key you could present an option this key file to
be saved separately as once the system is rebooted this key is gone and
there is no way back (no chance of restoring the backup once the key is
gone!).
I am glad to help you out with the encryption bit/issues if you need (I do
that sort of thing in my daily job) - just let me know. I am not certain if
you can see my email from my profile (to avoid cluttering the feedback
section in here unnecessarily), but we can continue this over email if you
need - let me know.
--
mr-4
[reply]
[top]
[»]
Re: /random and /urandom devices
by msetzerii - Jul 26th 2008 01:10:06
Thanks for the info. Didn't know that. It was just sitting there for a long
time, and nothing was happening. Then I tried it with the urandom, and it
was almost immediate. Could also been the method that they listed in the
readme to generate the key.
I did a test with a 500M file, and did notice one strange thing. Then
encrypted and decrypted version of the file where not the same size as the
original file, and diff shows the original and decrypted file as being
different.
The file was a zip file, and the decrypted file did pass the zip test. I
also noted that the encrypted and decrypted files were a multiple of 512
bytes, where the original file was about 300 bytes smaller.
Also, you need to enter the pass-phrase at default at run time, and not
clear on if that would be possible with the other piping, but looking at
options, that might be done with an option putting it in a temporary
file.
Also, there would have to be some way of getting the key file. Either
downloading it from an ftp site, or from some other device.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: /random and /urandom devices
by mr4 - Jul 26th 2008 02:32:30
Just sent you an email to the address on your profile with two text files
attached. Let me know if you haven't got it.
--
mr-4
[reply]
[top]
[»]
Dell Inspiron 1501
by Brian McDonald - Jul 8th 2008 23:16:52
Hello, I'd first like to say thank you for this awesome project.
Unfortunately though I've run into a problem, I am unable to load this
software properly on my laptop. No matter what option I choose it always
freezes at the line below:
ACPI: AC Adapter [ACAD] (on-line)
Any ideas?
--
--
Sincerely,
Brian McDonald
[reply]
[top]
[»]
Re: Dell Inspiron 1501
by msetzerii - Jul 9th 2008 01:45:44
Did you try using Kernel Option I:
That option has noacpi, which has been required for some notebooks to
work. I had been told by one user that the newer kernels didn't seem to
have the problem, but it might be some do, some don't.
If you have tried this, and other kernels, I'll have to look at it.
Are you using the 0.26 version?
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Release of 0.26 version - Shortly
by msetzerii - Jul 5th 2008 07:10:28
Getting ready to release 0.26, and here is pretty much what I have at the
moment.
ftp://amd64gcc.dyndns.org/g4l-v0.26.iso
ftp://amd64gcc.dyndns.org/g4l-v0.26.devel.tar.gz
Wasn't originally planning on releasing a version while I was in the
states on Summer Break, but a few users have had issues where the 0.26
alpha worked
The changes between 0.25 and 0.26 are:
Latest changes with 0.26a
added mdev as hotplug program
updated ntfs-3g
kernel 25.9
busybox 1.11.0 with patches to 7/02/2008
g4l30 script that shows partition size and type
added sfdisk and fsck programs.
added ifcheck script by dan.s@hostdime.com to recognize
multiple eth devices and activate the first that gets a
link for systems where eth0 is not the default.
Updated syslinux from 3.55 to 3.63
kernel 25.10 (added last minute untested)
The changes to the g4l script itself are minor, and mainly just the use of
the results of the ifcheck script to use the nic that gets an IP from dhcp
instead of defaulting to eth0. The busybox adn kernel updates are the
biggest. A little clean up on files, but did forget to change the 0.26a to
0.26 in the latest there, but that is all I can think of at the moment. If
no comments on these version by Monday, I will go thru the process of doing
the update on Sourceforge and Freshmeat, and it should shortly become the
official released version.
Thank You.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
g4l randomly stops during network back up
by Simon - Jul 2nd 2008 20:37:10
Hi!
I'm trying to use g4l to back up a 2 GB flash disk connected to the IDE
interface as hdc (with dma deactivated using the kernel paramter
ide=nodma). The flash disk is the main system disk, and I've booted from a
g4l cd through an attached USB cdrom drive.
I'm trying to back up to an ftp server (FileZilla Server) on my main
Windows machine, using raw-mode without compression (both machines are on a
local 100 Mbps network, with a router and a switch as the only other
equipment). However, the transfer seems to simply stop at random after a
while. It simply stops. It is still showing up as connected on the ftp
server, and I can cancel the operation with Ctrl-C.
Does anyone have a solution to this? - or maybe just an idea where the
problem could be?
[reply]
[top]
[»]
Re: g4l randomly stops during network back up
by msetzerii - Jul 2nd 2008 20:55:00
A couple of things to try?
1. Right before starting the Backup
Ctrl-Alt F2 to open another terminal
login with g4l
run top to watch what is happening.
Ctrl-Alt F1 to switch to original window
Start Backup
Ctrl-Alt F2 to monitor top, and see what is going on.
I would also suggest using lzop, since it is usually faster then no
compression. The network is generally slower than the drives, so it will be
faster if that is where the problem occurs. With a 100MB boot partition,
lzop would take 3 seconds, gzip was 6 seconds, and none was 10 seconds, and
bzip was 18 seconds.
2. You didn't mention exactly which version you are using, but another
user recently had an issue with a usb device, and the latest alpha version
seemed to fix it. Not sure if it was the kernel, or the busybox updates.
I'm currently 7 timezones from my machines, so making versions to isolate
what exactly fixed it isn't time effective.
ftp://amd64gcc.dyndns.org/g4l-v0.26alpha8.iso
Is the latest version, and will probable be released shortly, was planning
to wait till I get back, but that will not be until the 30th of July.
--
Computer Science Instructor
Guam Community College
[reply]
[top]
[»]
Re: g4l randomly stops during network back up
by Simon - Jul 2nd 2008 21:24:46
Thank you for the quick response:)
I'm using the newest release version - v0.25 from SourceForge, and the
default kernel (with the added ide=nodma parameter).
I tried using lzop, but got the same result as with no compression.
Looking at the output from top, all processes simply stop at 0% cpu
utilization.
I'll try out the new alpha though... Currently downloading...
[reply]
[top]
[»]
Re: g4l randomly stops during network back up
by Simon - Jul 2nd 2008 22:09:24
The new alpha version did the trick - now works perfectly! Thanks:)
|
