Best posts made by Quazz

This is a DCHP server issue. It nearly always happens when there are two DHCP servers fighting over who gets to deliver information.

Like Scott said, you need to alter some registry keys after an upgrade.

Also, check if it didn’t sneakily create some new user.

But overall, I’d recommend simply installing a clean image.

@Sozen /images/dev should be empty (bar .mntcheck) after a successful upload.

However, given that it is obviously not working properly, something else is up. Might be worthwhile to try and capture again and watch it closely.

Does it resize before capture? (if no, then you’re not using resizable image type)

Does it capture every partition?

Does it report any errors at any stage?

@rmmadden Try

REG IMPORT yourfile.reg

regedit is a bad tool for commandline, imo. REG is the way to go for command line registry stuff.

Regedit on commandline won’t notify of any possible errors, whereas REG might.

Could you update to the latest revision and try again?

@Andy-Nicolson Any BIOS updates available per chance?

Sounds like a bug on that end.

@plegrand You can upvote individual posts by clicking on the arrow next to ‘quote’

Only moderators can solve threads.

@Abuelika Alright, so it’s definitely something to do with the kernel.

Could you try adding pcie_aspm=off to kernel parameters at FOG General settings - FOG_KERNEL_ARGS ?

@Tony-Fiorentino Since you mention it’s old, it’s possible the drive has some bad sectors, causing the write to fail.

@sjensen

Then as Wayne said, make a new thread for it, sounds like your install is… special. Sudo being denied permission is silly.

The reason it’s failing is because you’re trying to export a drive that’s mounted from an external source. In a default installation, this won’t work. Some people have tried some (I think it was kernel) changes to allow for it, but I wouldn’t recommend it.

Given that you have a Windows Server 2012 you should be able to set it up as an NFS server directly, no need to use Samba sharing.

@RobTitian16 This script seems to have quite a few problems.

#!/bin/bash 

ceol='tput el';
machine='dmiedecode -s system-product-name'; #Gets machine model
machine="${machine%"${machine##*[![:space:]]}"}" #Removes trailing space
system64="/ntfs/Windows/SysWOW64/regedit.exe"; #Determine if it's 64 bit or not
if [ -f "$system64" ]; then
	setarch="x64"
else
	setarch="x86"
fi

dots "Preparing Drivers";

#Create the local folder on the imaged PC: 
mkdir /ntfs/Windows/DRV &>/dev/null;
echo -n "In Progress";

# Only use CAB files for the drivers. 
# Place the cab files for the specific system on the FOG Server:
# i.e. Model Latitude E5410, Windows 7 x64 image would be: 
# /fog/Drivers/Win7/Latitude E5410/x64 
cabextract -d /ntfs/Windows/DRV "/images/Drivers/${machine}"/*.CAB &>/dev/null; 

# Add the driver location on the PC to devicepath in Registry: 
regfile="/ntfs/Windows/System32/config/SOFTWARE"
key="\Microsoft\Windows\CurrentVersion\DevicePath"
devpath="%SystemRoot%\inf;%SystemRoot%\DRV";
reged -e "$regfile" &>/dev/null <<EOFREG
ed $key
$devpath
q
y
EOFREG
# Remove the "In Progress" and replace it with: "Done"
echo -e "\b\b\b\b\b\b\b\b\b\b\b${ceol}Done"; 

Changed to use bash (as sh is more restrictive and does not support some of the things done here)

Changed =f if condition to the correct -f

setarch appears unused, however, so not even sure why that’s being checked for?

@RobTitian16 I don’t know what else you have in your scripts, I can only make guesses.

This script does not mount anything, so it wouldn’t be able to make the C:\Windows\DRV directory if the target is not mounted to begin with.

Of course if you have mounted it in another script (and not have it unmount it) then it should work, assuming everything works as expected.

option bootfile-name “undionly.kpxe”;

Is incorrect

option bootfile-name "undionly.kpxe";

Is correct.

Fix your quotation marks (perhaps elsewhere too!) and it should work fine after restarting the dhcp server.

@imagingmaster21 When FOG tries to create an image from the host it creates small text files with information about the partitions that are used during deployment so that it knows which to resize and which not, how big they should be etc.

You will find them typically in /images/IMAGENAME

The problem in your case is that your main data partition is locked in between two other partitions. As such the third partition starts at the marker for around 49.37GB or so (give or take). This means that the maximum size of the main partition will be limited by it.

On a default installation, the Windows partition itself is the last one

You’ll have to fix this on the host and recapture the image.

What OS and version are you running?

What does curl https://github.com --cipher rsa_aes_128_sha get you.

Curl error 35 is:

"Cannot communicate securely with peer: no common encryption algorithm(s)."

The local and remote systems share no cipher suites in common. This can be due to a misconfiguration at either end. It can be due to a server being misconfigured to use a non-RSA certificate with the RSA key exchange algorithm.

Also try curl https://github.com -v it should give us more info

I recall certain SSDs getting inexplicably low speeds.

Here’s the discussion about it on the partclone github page:

https://github.com/Thomas-Tsai/partclone/issues/112

Currently not solved and obviously a pretty big problem if you’re experiencing this.

Only info I’ve found myself is about 32 bit kernels on systems with 4GB+ RAM causing this issue.

Some older topics suggest the DMA is sometimes disabled on SSDs for some reason.

@george1421 The problem is for new kernels we need to update the kernel headers. Programs built against that (such as the programs in the init files) require the minimum supported kernel of those headers to run.

In practice, it will often work anyway, but sometimes the changes make it impossible, nothing that can really be done about that from our side afaik.

@BSZAdmin Your export file is wrong, /datastore/images/dev needs rw permissions, not ro.

In fact, you seem to have copied over the /images line for it, rather than /images/dev

Fix /etc/fstab, and re export

@jeremyvdv said in Mount NFS:

I made a redirection of the images folder of my debian to a nas.

What do you mean by that?

If you have mounted the NAS folder to your Debian machine then it won’t work since the Debian machine can’t “export” a folder that isn’t its own. (you can’t daisy chain it)

If you want to use a NAS as storage, you can try configuring it as a storage node.