FOG restore only one partition

[Tom Elliott]

@plegrand there can be a solution but I have to figure out what that solution will be.

[plegrand]

@Tom-Elliott
for the moment i reinstalled the station and i continu my tests. I’ve got one question :

I saw on the wiki that there is 4 images types :

Single Partition
Multiple Partition - Single Disk
Multiple Partition - All Disks
Raw Image

On the web frontend
Single Disk - Resizable - (1)
Multiple Partition Image - Single Disk (Not Resizable) - (2)
Multiple Partition Image - All Disks (Not Resizable) - (3)
Raw Image (Sector By Sector, DD, Slow) - (4)

As you can see the first option is different.
What is the good one
Is it single disk or single partition ?

And finally, instead opening a new thread, i’ve got a last question :
all my host have a red circle with exclamation point and this error message :
“no such device or address”. What does that mean ?

Thanks for your help

[plegrand]

@Tom-Elliott
Hello
i just made a test today with an other station
Dell Optiplex 760 with again 3 partitions
Here is the image definition :
default: Windows 2000/XP
Single Disk - Resizable
Everything

The capture seems to works fine and i can see in the “images” directory :

ls -al /home/images/Optiplex760780/
total 8355044
drwxrwxrwx 2 root root 4096 févr. 23 10:09 .
drwxrwxrwx 7 root root 4096 févr. 23 10:09 …
-rwxrwxrwx 1 root root 3 févr. 23 09:44 d1.fixed_size_partitions
-rwxrwxrwx 1 root root 249 févr. 23 09:49 d1.minimum.partitions
-rwxrwxrwx 1 root root 30 févr. 23 09:49 d1.original.fstypes
-rwxrwxrwx 1 root root 0 févr. 23 09:44 d1.original.swapuuids
-rwxrwxrwx 1 root root 3706587 févr. 23 09:49 d1p1.img
-rwxrwxrwx 1 root root 8551432512 févr. 23 10:09 d1p2.img
-rwxrwxrwx 1 root root 392584 févr. 23 10:09 d1p3.img
-rwxrwxrwx 1 root root 249 févr. 23 09:44 d1.partitions

But when i try to deploy this image on the same station i’ve got again this error message at the beginning :
“This is not a partclone image”

Now the station is “dead” and cant boot “Error loading operating system”

It’s a real problem because we have a lot of this kind of station

[Tom Elliott]

I just added some more code and I think I finally found the problem.

[plegrand]

@Tom-Elliott Hello
does that mean that with the latest version it will work ?
Do you want i try ?
Thanks for your help

[Sebastian Roth]

@plegrand Yes, would be great if you can upgrade to the latest version and see if it’s fixed. Please report back!

[plegrand]

@Sebastian-Roth
i have to capture again before deploy ?

Just to be clearer, here is my configuration :

On my own pxe server(dhcp and tftp) 192.168.151.247
dhcp is configured like that :

 next-server 192.168.151.247;
 filename "pxelinux.0";

then on my tftp server i’ve got pxelinux.0 into the tftp root directory
and
this configuration

LABEL FOG_ETUDIANTS
KERNEL fog/ipxe.krn dhcp && chain http://192.168.39.243/fog/service/ipxe/boot.php?mac=${net0/mac}

192.168.39.243 is the fog ip address
Do i have to put on my tftp server the new ipxe.krn file ?

[Tom Elliott]

@plegrand you should not have to re upload the image.

[plegrand]

@Tom-Elliott Hello
Fog 6427 / 4874

here is the tests i made

• Restore the station with an old symantec ghost image
• List partitions with diskpart

DISKPART> list disk
 Disque ###  Statut      Taille     Libre     Dyn  Gpt
 --------  ---------  -----------  --------   ---  ---
 Disque 0    Connecté     149 GB      0 B

 DISKPART> select disk 0
 Le disque 0 est maintenant le disque sélectionné.

 DISKPART> list partition
 Partition ###  Type              Taille     Décalage
 -------------  ----------------  -------  -----------
 Partition 1    OEM                126 MB    20 KB
 Partition 2    Principale          49 GB   126 MB
 Partition 3    Principale         100 GB    49 GB

• Defragmentation of all the partition
• On fog create a image :

	default: Windows 2000/XP
	Single Disk - Resizable
	Everything

• Assign the new image to the host
• On fog create a new capture task for the station as a debug task
• Launch the task
• Capture works fine
  On the server fog :

	d1.fixed_size_partitions
	d1.minimum.partitions
	d1.original.fstypes
	d1.original.swapuuids
	d1p1.img
	d1p2.img
	d1p3.img
	d1.partitions

	cat d1.fixed_size_partitions 
	1:

	cat d1.minimum.partitions 
	label: dos
	label-id: 0x2bd2c32a
	device: /dev/sda
	unit: sectors

	/dev/sda1 : start=          40, size=      257480, type=de
	/dev/sda2 : start=      257520, size=    31386360, type=7, bootable
	/dev/sda3 : start=   102657680, size=      779520, type=7

	cat d1.original.fstypes 
	/dev/sda2 ntfs
	/dev/sda3 ntfs

	cat d1.partitions 
	label: dos
	label-id: 0x2bd2c32a
	device: /dev/sda
	unit: sectors

	/dev/sda1 : start=          40, size=      257480, type=de
	/dev/sda2 : start=      257520, size=   102400160, type=7, bootable
	/dev/sda3 : start=   102657680, size=   209839360, type=7

At reboot, the station stay displaying “Booting…”

Then i try to deploy the image

• Create a new deploy task for the host

During the restauration of /dev/sda1, i’ve got the following error message from partclone

Target partition size (132MB) is smaller than source 132 MB. 
Use option -C to disable size checking (Dangerous)

And just after

Image failed to restore and exited with exit code 0 (writeImage)
Args passed /images/test/d1p1 img* /dev/sda1

Then the only solution is to redeploy the old ghost image

[Tom Elliott]

@plegrand have you tried after updating to the latest.

[plegrand]

@Tom-Elliott
no
do i have to capture again or just deploy ?

[plegrand]

@Tom-Elliott hello
i just made the following test with latest version : 6476 / 4895
just create deploy task
same problem

After upgrade, Do i have to put on my own tftp server the new ipxe.krn file ?

[Sebastian Roth]

@plegrand As Tom said, could you please upgrade to the latest version. He is saying that you don’t need to re-upload. But I’d say it wouldn’t hurt anyway.

You don’t need to update ipxe.krn on your TFTP server if booting the client works fine. The iPXE binary is not related to the error you see AFAIK!

[plegrand]

@Sebastian-Roth Same problem

[Sebastian Roth]

@plegrand Please boot that client you tried to write the image to in debug deploy again (schedule a normal deploy task but tick “Schedule task as a debug task” in the advanced settings box just before you hit the button “CREATE DEPLOY TASK FOR …”).

When you get to the shell please run sfdisk -d /dev/sda and post a picture of the output here (we need the exact numbers!!).

[plegrand]

@Sebastian-Roth

label: dos
label-id: 0x2bd2c32a
device: /dev/sda
unit: sectors

/dev/sda1 : start = 63,         size = 257024,    type = de
/dev/sda2 : start = 257087,     size = 102401024, type = 7, Bootable
/dev/sda3 : start = 102658111,  size = 209841152, type = 7

[Sebastian Roth]

@plegrand Did you copy&paste the numbers or typed them from the screen? I hope there is no typo in them!

If the figures are correct this is simple maths. The original first partition (as seen further down) starts at sector 40 and ends at 257480. Makes 257440 in total. After writing to the new disk the first partition starts at 63 (which should not be an issue) but ends at 257024. Makes a total of 256961 sectors for the first partition. Slightly smaller than the original one.

I guess our re-sizing logic is still hating those recovery partitions (type=de)!

[Sebastian Roth]

I think I found where the issue is. Please give me some more time to see why things go wrong…

[Sebastian Roth]

@plegrand Can you please do a debug UPLOAD task on the original client you pull the image from. You don’t need to do the full upload. I just need to get some information. Run the following command when you get to the shell and post the full output here: blkid -po udev /dev/sda1

You can simply delete the task in the web interface and shutdown that client after you got the information.

[plegrand]

@Sebastian-Roth

ID_PART_ENTRY_SCHEME=dos
ID_PART_ENTRY_UUID=2bd2c32a-01
ID_PART_ENTRY_TYPE=0xde
ID_PART_ENTRY_NUMBER=1
ID_PART_ENTRY_OFFSET=63
ID_PART_ENTRY_SIZE=257024
ID_PART_ENTRY_DISK=8:0