ecmastrean Absent Member.
Absent Member.
2977 views

Imaging Failure With Traditional Hard Drives

We have seen Dell computers (newer models such as OptiPlex 5040 and 5050 with sixth and seventh gen Intel CPU's) inconsistently fail to image to Windows 10 x64. The BIOS is in UEFI mode and the imaging fails at the very beginning of the image. It may fail 4 times out of 5. This only seems to happen with traditional spinning hard drives. We are seeing this error message at the bottom of the screen when it fails:
bin/zenAdvancedScript: line 144: 2997 Floating point exceptioning rp $IMGSVR 1703Win10.zmg
Loading scripted installers

Here is a portion of our imaging script:

Line 143 echo "Loading Image"
Line 144 img rp $IMGSVR 1703Win10.zmg
Line 145
Line 146 echo "Loading scripted installers"
Line 147 img rp $IMGSVR 1703addon.zmg
Line 148
Line 149 if [ $drivers ]; then
Line 150 img rp $IMGSVR $drivers.zmg
Line 151 fi

When the 1703Win10.zmg fails (this is the OS) it continues to load the 1703addon.zmg (antivirus, ZENworks installer, etc) then the drivers. The addon and driver image copy OK, it only fails when copying 1703Win10.zmg to the computer.

We seen this behavior on ZENworks 11.4.1 then on ZENworks 2017. Has anyone seen this behavior before or have any ideas?
Labels (1)
Tags (3)
0 Likes
13 Replies
Not applicable

Re: Imaging Failure With Traditional Hard Drives

I moved the thread to a more appropriate forum area.

Are you using sysprep? Did you run chkdsk, defrag?
0 Likes
ecmastrean Absent Member.
Absent Member.

Re: Imaging Failure With Traditional Hard Drives

Yes, we are using sysprep. We have ran check disk, defrag and ran diagnostics on the drives and they pass. Do you mean run check disk and defrag before we make the OS image?
0 Likes
Micro Focus Expert
Micro Focus Expert

Re: Imaging Failure With Traditional Hard Drives

What is the KERNEL Version being loaded during PXE Preboot?
0 Likes
Not applicable

Re: Imaging Failure With Traditional Hard Drives

Yes, it's good to run chkdsk and defrag before image creation.
0 Likes
ecmastrean Absent Member.
Absent Member.

Re: Imaging Failure With Traditional Hard Drives

We are using PXE agent 17.0.0 (ZENworks 2017 with the 10 March 2017 driver update installed). We run check disk and defrag before image creation.

We are using UEFI BIOS with Windows 10 built on the GPT file system, so our image creates 4 partitions:
1. Recovery for GPT
2. EFI BIOS
3. Blank
4. Image Partition

We had problems installing drivers so when we used the ZENworks image explorer to look at the image, it installed the OS to the 4th partition, so we built the addons and driver packs to install on the 4th partition specifically. Since we do not have problems installing the addons or drivers, just the image, it may be because of where the image is writing to. Also the addons and drivers are not over 2GB each, but the image is around 5 GB. So it may be a file size problem too since the image starts to write to the disk then it fails before 1% complete.

If we take a traditional hard drive and delete all partitions manually then image it, the computer images fine.

Here are our commands delete partitions and image:
echo "Deleting existing Partitions:"
img pd-all;

echo "Loading Image"
img rp $IMGSVR 1703Win10.zmg

echo "Loading scripted installers"
img rp $IMGSVR 1703addon.zmg

if [ $drivers ]; then
img rp $IMGSVR $drivers.zmg

We copied the img pd-all; or img rp commands from a script used to install Windows 7 using NTFS and legacy BIOS. If we can find a command that deletes all partitions then that should solve the problem.
0 Likes
Micro Focus Expert
Micro Focus Expert

Re: Imaging Failure With Traditional Hard Drives

You may be missing a space in your script......

img pd-all; should be img pd -all
0 Likes
Micro Focus Expert
Micro Focus Expert

Re: Imaging Failure With Traditional Hard Drives

that being said..........i doubt that is your issue.......I suspect the issue the is KERNEL Version being loaded...........Every Imaging Update has 2 Kernel versions.
Which Version are those Dells loading?
0 Likes
ecmastrean Absent Member.
Absent Member.

Re: Imaging Failure With Traditional Hard Drives

We noticed the missing space in the command img pd -all. However we also noticed there was a dash in front of pd in the documentation so we changed it to: img -pd -all; We imaged a problem computer last night and it imaged OK, but it has been my experience that this problem is intermittent, so we may have got lucky. We let our techs know about the change and will see if they can image without problems today. I will post an update tomorrow. I found two versions for the PXE kernel, I'm not sure which one you need:
ELILO 3.14
EFI 2.40
0 Likes
Micro Focus Expert
Micro Focus Expert

Re: Imaging Failure With Traditional Hard Drives

https://download.microfocus.com/Download?buildid=JBf94I6uK7c~
New Linux Kernel versions in this Imaging Engine update:
Kernel version for 32 bit: 3.0.101-63-default
Kernel version for 64 bit: 4.4.21-69-default

Normally the 64 bit kernel will load on devices that support UEFI (Even in Emulation)
The 32bit Kernel on BIOS devices.

The 64 bit is newer and may be needed to properly support newer hardware.

There is one file in particular that helps determine which one loads that sometimes does not update properly causing the older kernel to load on newer devices.
The updating issue on that file should be fixed in the July imaging update due any day now.....

So if you are seeing the older 3.0 Kernel Loading, the issue may be resolved in the newer 4.4 Kernel....especially if the issue is seen mostly on newer hardware.
0 Likes
ecmastrean Absent Member.
Absent Member.

Re: Imaging Failure With Traditional Hard Drives

We have the ZCM 2017 Imaging Driver Update for March 2017 driver update installed. Since the 3.14 version is loading, you're thinking it is the 32 bit version and not 64? I'm waiting for the July update too since it may solve some other issues we hare having.

By the way, we are still having problems imaging so changing the formatting command to img -pd -all did not help. I hope the July update will fix it.
0 Likes
ecmastrean Absent Member.
Absent Member.

Re: Imaging Failure With Traditional Hard Drives

We manually imaged 16 OptiPlex 5040 computers in a lab yesterday. My manually I mean booting into maintenance mode and then typing img then pressing F6 to manually select the server and imaging file to use. I wanted to rule out any problem with our script. Seven of the 16 computers imaged OK, the other 9 got the error about a floating point exception error. What is interesting is that all the computers that failed did so copying the same file: EFI\Boot\BOOTX64.EFI. If we would reboot and try to image the computers that failed, it may take a few more tries to image them. This was at the end of the day when there was not much network traffic. Each of our sites has a local server used for imaging, so when we imaged 20 OptiPlex 5050's at another location, 7 of them imaged ok and 13 failed. So the fact that 7 seems to be the limit of how many will image OK at a time seems to point to a network issue, but we only have this problem on newer Dell desktops (OptiPlex 5040, 5050 and Precision T5810) using UEFI BIOS and Windows 10 with GPT partition tables. Dell laptops and Windows 7 computers using MBR and Legacy BIOS do not have this problem. So it makes me think it is a computer specific problem, or maybe a combination of network and specific computer problem.
0 Likes
kirchke Absent Member.
Absent Member.

Re: Imaging Failure With Traditional Hard Drives

ecmastrean;2462215 wrote:
We manually imaged 16 OptiPlex 5040 computers in a lab yesterday. My manually I mean booting into maintenance mode and then typing img then pressing F6 to manually select the server and imaging file to use. I wanted to rule out any problem with our script. Seven of the 16 computers imaged OK, the other 9 got the error about a floating point exception error. What is interesting is that all the computers that failed did so copying the same file: EFI\Boot\BOOTX64.EFI. If we would reboot and try to image the computers that failed, it may take a few more tries to image them. This was at the end of the day when there was not much network traffic. Each of our sites has a local server used for imaging, so when we imaged 20 OptiPlex 5050's at another location, 7 of them imaged ok and 13 failed. So the fact that 7 seems to be the limit of how many will image OK at a time seems to point to a network issue, but we only have this problem on newer Dell desktops (OptiPlex 5040, 5050 and Precision T5810) using UEFI BIOS and Windows 10 with GPT partition tables. Dell laptops and Windows 7 computers using MBR and Legacy BIOS do not have this problem. So it makes me think it is a computer specific problem, or maybe a combination of network and specific computer problem.


I wanted to chime in and say that I am experiencing the very same problem especially with your Dell OptiPlex 7440 AIO machines. Our master image will sometimes cause a floating point error, which has been described, when trying to restore BOOTX64.EFI. Other machines that are completely identical in terms of hardware and master image being placed onto them have no issue. We seem to have better luck getting it to work within Maintenance Mode and also if we fiddle with the computer's firmware (although there appears to be no way to accurately reproduce the firmware setting changes to get it to work consistently - it could just be that multiple attempts at imaging just work).

As mentioned above, I can get the device to image most of the time if I change around the following firmware settings:
- Disable/Enable legacy boot options
- Change SATA operation from RAID to AHCI or vise versa
- Change boot order
- Reset BIOS to factory defaults
- Update firmware (usually does not fix the problem on its own)

I've also imaged machines with our EFI image that have worked the first time and will throw a floating point error if I attempt to reimage them.

ZCM Version: 2017 Update 1
Imaging Operating System: SUSE Linux Enterprise 12 SP2
Kernel Version: 4.4.21-69
Operating System attempting to be restored: Windows 10 x64-bit Creators Update
0 Likes
ecmastrean Absent Member.
Absent Member.

Re: Imaging Failure With Traditional Hard Drives

Thanks kirchke for your findings. We tried to image them multiple times (manually and by our script) and it eventually worked. Last night I applied the ZCM 2017 Imaging Driver Update for July 2017 and after that we have not had any problems imaging the OptiPlex 5050. We imaged 17 this morning at a time, whereas before, only 7 out of a given number would image. We tired to image the OptiPlex 5050 we have in our office for testing and it imaged OK without giving the floating point error. So install the July imaging driver update and see if that fixes your problem. We are using the following:

ZCM Version: 2017
Imaging Operating System: SUSE Linux Enterprise 12 SP2
Operating System attempting to be restored: Windows 10 x64-bit Edu
BIOS: EFI
OS File System: GPT
0 Likes
The opinions expressed above are the personal opinions of the authors, not of Micro Focus. By using this site, you accept the Terms of Use and Rules of Participation. Certain versions of content ("Material") accessible here may contain branding from Hewlett-Packard Company (now HP Inc.) and Hewlett Packard Enterprise Company. As of September 1, 2017, the Material is now offered by Micro Focus, a separately owned and operated company. Any reference to the HP and Hewlett Packard Enterprise/HPE marks is historical in nature, and the HP and Hewlett Packard Enterprise/HPE marks are the property of their respective owners.