Partclone suggests you should have 1548900463201413591 bytes memory

[Michael Thomas]

Hello

uploading an image seemed to have worked, but download never had success. i don’t know where could find any
detailed error message. Do i have to set up a syslog server? Where to go? Any hints, please?

More detail:
fog 1.2 on ubuntu
Linux fog2 3.2.0-67-generic-pae #101-Ubuntu SMP Tue Jul 15 18:04:54 UTC 2014 i686 i686 i386 GNU/Linu

As i only see “flashings” of PXE i can’t read, I tried using the “debug download”
and see a funny error message:

i mount the nfs share from the fog server
#mkdir /images
#mount -o nolock nn.nn.nn.nn://images/ /images

i coppied the d1.mbr to /dev/sda
#dd if=/images/W7/d1.mbr of=/dev/sda

#/usr/sbin/partclone.ntfs -s /images/W7/d1p1.img -o /dev/sda1 -r
Partclone v0.2.69 [url]http://partclone.org[/url]
Starting to restore image (/images/W7/d1p1.img) to device (/dev/sda1)
There is not enough free memory, partclone suggests you should have 1548900463201413591 bytes memory
Partclone fail, please check /var/log/partclone.log !

Pretty clear error message but: Where do i get 1.548.900.463.201.413.591 of RAM? From NSA?

I tried on the fog server to see if the image is OK at least:

root@fog2:~# /home/fog/fog_1.2.0/src/buildroot/package/fog/partclone32/partclone.chkimg -s /images/W7/d1p1.img
Partclone v0.2.69 [url]http://partclone.org[/url]
Starting to check image (/images/W7/d1p1.img)
There is not enough free memory, partclone suggests you should have 1548900463201413591 bytes memory
Partclone fail, please check /var/log/partclone.log !

(On both machines i got the same memory value)

i installed the ubuntu partclone package on the fog server:

root@fog2:~# partclone.chkimg -s /images/W7/d1p1.img
Partclone v0.2.43 [url]http://partclone.org[/url]
Starting to check image (/images/W7/d1p1.img)
The Image magic error. This file is NOT partclone Image
Partclone fail, please check /var/log/partclone.log !

root@fog2:~# file /images/W7/d1p1.img
/images/W7/d1p1.img: gzip compressed data, from Unix, last modified: Thu Jul 31 15:42:20 2014, max compression

What do i miss?
I have no clue where to search an waht’s going wrong.

root@fog2:~# ll /images/W7/
total 4172508
drwxrwxrwx 2 root root 4096 Jul 31 13:52 ./
drwxrwxrwx 6 root root 4096 Jul 31 13:53 …/
-rwxrwxrwx 1 root root 512 Jul 31 13:42 d1.mbr*
-rwxrwxrwx 1 root root 8274330 Jul 31 13:42 d1p1.img*
-rwxrwxrwx 1 root root 4136221212 Jul 31 13:52 d1p2.img*
-rwxrwxrwx 1 root root 512 Jul 31 13:52 d2.mbr*
-rwxrwxrwx 1 root root 128126142 Jul 31 13:53 d2p1.img*

fog server i a ubuntu server running in Oracel Virtual Box 4.3.15 r95226 (4.3.14 is broken)
client is a virtual machine Oracel Virtual Box 4.3.15 r95226 too.

[Michael Thomas]

Append:

i previouly to the above text i entered the “fog” command in the debug console.
It gives a lot of status informations, all seems to be well,
but when i entered the “last” return, the client closes immediately. I saw no partclone dialog, only the reoboot screen.
No hint what was going wrong.

Now, after manually mounting the nfs share and manually overwirting(dd) the mbr and
got those rediculose error messages, i lanches the fog command (i thoug it would be senseless too)
It starts and
I could read the line similar to “deleting old start mbr”
and then partclone starts to run!

What’s going on there?
Do i have to wipe(zero out) the disk manually first? (i only created an empty vdi)
Would not make sense for a disk cloning software to do the manually on the cloent, or?

[Tom Elliott]

I think the first post is due to how you’re resending the image to the disk:
[code]partclone.restore -s /images/W7/d1p1.img -O /dev/sda1[/code]

This won’t work though, simply because the image file is gzipped. So we pass it properly with:
[code]cat /images/W7/d1p1.img | gunzip -d -c | partclone.restore -O /dev/sda1 -N -f 1[/code]

[Michael Thomas]

Thanks for the fast reply(during weekend i had no access). I wasn’t aware that a have to unzip the img-file and
that partclone does not use any “magic number” to detect the .img format.
OK, that explains why may manually restore did not work neither.
But i tried that only because the normal download way falied without any hint what’s going wrong (The task is clearly removed from schedule as if all went well…that surprised me.)

I inspected the logs files i know (in /var/log/) but didn’t find a hint. So my idea was that fog is logging somewhere else using the syslog protocol. But there are no logs fog at all?

---

spec_
FOG Version 1.2
Linux fog2 3.2.0-67-generic-pae #101-Ubuntu SMP Tue Jul 15 18:04:54 UTC 2014 i686 i686 i386 GNU/Linu
Running Virtual Box 3.15 2 CPU 90% PIIX3, IO-APIC, no EFI, 512MB RAM, VT/x+Nested

I don’t change the DHCP server settings (066,067 unset).
Instead I made and boot an iPXE-Iso which includes the required Server/File settings and PXE modules fog needs.
So i don’t have to finger in the BIOS enabling PXE and always run a known working PXE (there were problems with VirtualBOX-BIOS-PXE some time ago)

IiPXE.iso[/I]
I#!ipxe[/I]

Idhcp[/I]
Iset filename undionly.kpxe[/I]
Iset filename default.ipxe[/I]
Iset next-server 10.51.251.180[/I]

Iecho Z My IP address: ${net0/ip}[/I]
Iecho Z Subnet mask: ${net0/netmask}[/I]
Iecho Z Will boot ${filename} from ${next-server}[/I]

Iprompt --key 0x02 --timeout 20 Press Ctrl-B for the iPXE command line… && shell ||[/I]

I# chain tftp://10.51.250.180/undionly.kpxe[/I]
Ichain tftp://${next-server}/${filename}[/I]

Iautoboot[/I]

I---------------[/I]
Idefault.ipxe[/I]
I# cat /tftpboot/default.ipxe[/I]
I#!ipxe[/I]
Iecho This is tftp://default.ipxe[/I]
Icpuid --ext 29 && set arch x86_64 || set arch i386[/I]
Iecho CPU-Id is ${arch} ${net0/mac} ${net1/mac} ${net2/mac}[/I]
Iparams[/I]
Iparam mac0 ${net0/mac}[/I]
Iparam arch ${arch}[/I]
Iisset ${net1/mac} && param mac1 ${net1/mac} || goto bootme[/I]
Iisset ${net2/mac} && param mac2 ${net2/mac} || goto bootme[/I]

Iecho Note: default.ipxe chain [url]http://10.51.251.180/fog/service/ipxe/boot.php##params[/url][/I]
Iecho params param ##params[/I]
I#shell[/I]
Ichain [url]http://10.51.251.180/fog/service/ipxe/boot.php##params[/url][/I]

Current progress:
I can enter the fog management server
I can boot PXE-Iso
http:boopt.php is lanched and that iPXE-menu is shown
i can register the box, run memtest etc.
all well and flawless.
i can upload (i see partcones menu, it shows it is uploading bar and speed (approx 12MB/s in entire virtual environment)
On the server i can add a task to download to an other, fresh VM with a fresh VDI.
The PXE boot occurs, it checks the settings, does not find a problem, and…closes instead of downloading.

[SIZE=5]Idea:[/SIZE]
[SIZE=5]As if partclone does not make any plausibility tests to the image, may i assume it does not validate the MBR too in any way too?[/SIZE]
[SIZE=5]So if the MBR/Partition contains garbage, it silently an imediately dies? (Would explain why i don’t see any logs and that it is now working, after i manually “initialize” the MBR/partion?[/SIZE]

[SIZE=5]Partclone seems to die before it can zero out the mbr…or?[/SIZE]

[SIZE=5]Question:[/SIZE]
[SIZE=5]Is my assumption right partclone died as MBR was filled with garbage (random noise)?[/SIZE]
[SIZE=5]If my assumptions are right:
Will partclone get input validations so that such problems will be logable with comprehensive error messages?[/SIZE]
[SIZE=5]Until that day:[/SIZE]
[SIZE=5]Is there a way to “short” zero out the MBR before partcone runs? (i don’t want to wipe the entire 4TB disk :-))
I see, that the fog script is attempting this, but i think it’s too late:
something already died on the non-sense MBR.
Any help?[/SIZE]

[Michael Thomas]

[quote=“Michael Thomas, post: 34689, member: 25383”]Thanks for the fast reply(during weekend i had no access). I wasn’t aware that a have to unzip the img-file and
that partclone does not use any “magic number” to detect the .img format.
OK, that explains why may manually restore did not work neither.
But i tried that only because the normal download way falied without any hint what’s going wrong (The task is clearly removed from schedule as if all went well…that surprised me.)

I inspected the logs files i know (in /var/log/) but didn’t find a hint. So my idea was that fog is logging somewhere else using the syslog protocol. But there are no logs fog at all?

---

spec_
FOG Version 1.2
Linux fog2 3.2.0-67-generic-pae #101-Ubuntu SMP Tue Jul 15 18:04:54 UTC 2014 i686 i686 i386 GNU/Linu
Running Virtual Box 3.15 2 CPU 90% PIIX3, IO-APIC, no EFI, 512MB RAM, VT/x+Nested

I don’t change the DHCP server settings (066,067 unset).
Instead I made and boot an iPXE-Iso which includes the required Server/File settings and PXE modules fog needs.
So i don’t have to finger in the BIOS enabling PXE and always run a known working PXE (there were problems with VirtualBOX-BIOS-PXE some time ago)

IiPXE.iso[/I]
I#!ipxe[/I]

Idhcp[/I]
Iset filename undionly.kpxe[/I]
Iset filename default.ipxe[/I]
Iset next-server 10.51.251.180[/I]

Iecho Z My IP address: ${net0/ip}[/I]
Iecho Z Subnet mask: ${net0/netmask}[/I]
Iecho Z Will boot ${filename} from ${next-server}[/I]

Iprompt --key 0x02 --timeout 20 Press Ctrl-B for the iPXE command line… && shell ||[/I]

I# chain tftp://10.51.250.180/undionly.kpxe[/I]
Ichain tftp://${next-server}/${filename}[/I]

Iautoboot[/I]

I---------------[/I]
Idefault.ipxe[/I]
I# cat /tftpboot/default.ipxe[/I]
I#!ipxe[/I]
Iecho This is tftp://default.ipxe[/I]
Icpuid --ext 29 && set arch x86_64 || set arch i386[/I]
Iecho CPU-Id is ${arch} ${net0/mac} ${net1/mac} ${net2/mac}[/I]
Iparams[/I]
Iparam mac0 ${net0/mac}[/I]
Iparam arch ${arch}[/I]
Iisset ${net1/mac} && param mac1 ${net1/mac} || goto bootme[/I]
Iisset ${net2/mac} && param mac2 ${net2/mac} || goto bootme[/I]

Iecho Note: default.ipxe chain [url]http://10.51.251.180/fog/service/ipxe/boot.php##params[/url][/I]
Iecho params param ##params[/I]
I#shell[/I]
Ichain [url]http://10.51.251.180/fog/service/ipxe/boot.php##params[/url][/I]

Current progress:
I can enter the fog management server
I can boot PXE-Iso
http:boopt.php is lanched and that iPXE-menu is shown
i can register the box, run memtest etc.
all well and flawless.
i can upload (i see partcones menu, it shows it is uploading bar and speed (approx 12MB/s in entire virtual environment)
On the server i can add a task to download to an other, fresh VM with a fresh VDI.
The PXE boot occurs, it checks the settings, does not find a problem, and…closes instead of downloading.

[SIZE=5]Idea:[/SIZE]
[SIZE=5]As if partclone does not make any plausibility tests to the image, may i assume it does not validate the MBR too in any way too?[/SIZE]
[SIZE=5]So if the MBR/Partition contains garbage, it silently an imediately dies? (Would explain why i don’t see any logs and that it is now working, after i manually “initialize” the MBR/partion?[/SIZE]

[SIZE=5]Partclone seems to die before it can zero out the mbr…or?[/SIZE]

[SIZE=5]Question:[/SIZE]
[SIZE=5]Is my assumption right partclone died as MBR was filled with garbage (random noise)?[/SIZE]
[SIZE=5]If my assumptions are right:[/SIZE]
[SIZE=5]Will partclone get input validations so that such problems will be logable with comprehensive error messages?[/SIZE]
[SIZE=5]Until that day:[/SIZE]
[SIZE=5]Is there a way to “short” zero out the MBR before partcone runs? (i don’t want to wipe the entire 4TB disk :))[/SIZE]
[SIZE=5]I see, that the fog script is attempting this, but i think it’s too late: [/SIZE]
[SIZE=5]something already died on the non-sense MBR.[/SIZE]
[SIZE=5]Any help?[/SIZE][/quote]

[Junkhacker]

is there a reason you quoted yourself without adding anything new?

[Michael Thomas]

Hi
new try:

generated a new 25GB dynamical disk in virtual box
select this disc in the VM.
tried a dowload: fails as always
(so the problem is reproducible for me, but why am i the only one?)
tried debug download:
did hexdump -C -n 32M /dev/sda
hd shows all zeroes (not random numbers)
“fdisk -l” shows the right dimensions
fdisk /dev/sda complaints “Devie does not contain a recognized partiotion table”
“Created a new DOS disklabel with disk identfier …”
“w” write does not alter the disk…?
Where did fdisk -l got its data from?

tried “sh -x fog” and initiate a disk surface test:
Finds /dev/sda and finishes in fraction of a second
(would like to add a screendump, there seems to be no way with virtual box, sorry)

#fogpartinfo --list-parts /dev/sda
Error: /dev/sda: unrecognized disk label

#fogpartinfo --list-devices
Error: /dev/sda: unrecognized disk label

???

[Michael Thomas]

[quote=“Junkhacker, post: 34699, member: 21583”]is there a reason you quoted yourself without adding anything new?[/quote]
No, a quote was not my intention, sorry.
sorry this BB is new to. I’ll try to find the button to turn that useless quoting off.