Posts made by Polar Bear

Hello,

I’ve taken simple approach of using USB NIC (Realtek 8152) with ipxe added to flash drive as EFI boot module.

The best result achieved with ‘ecm’ compilation option for boot module

make bin-x86_64-efi/ecm.usb
dd if=bin-x86_64-efi/ecm.usb of=/dev/{usb_drive}

Once configuration was added into DHCP/DNS server the boot process not much different as if NIC was integrated into laptop and sit on PCI bus.

A few FOG imaging test gave same speed result as if NIC was integrated into the laptop ( for test was used HP Elitebook Folio 1040 G2, at this moment a laptop without wired NIC is not available).

Thanks to everyone who responded to my post.

Hello,

nowadays some laptops do not have ethernet port for wired connection. The only available option is USB NIC.

How such situation can be resolved?
What would be a procedure?

My understanding is that USB drive with linux kernel and driver for USB NIC interface would require to initiate DHCP request to communicate with FOG server.

Thank you for any leads to helpful information.

@george1421

Thank you George,

I only can say ‘Busy Busy Busy’ as Twiki – I need to absorb the information.

When I mentioned ‘the installation continues as from local media (DVD)’ – I meant that as files was made available from network resource the process will continue in same way as if it was run from local media (I meant that process will not be different, except that files be accessed from network and not from local computer media).

Thank you again George
Polar Bear

@george1421

Thank you George,

your explanation cleared some my questions but not all of them (I was not ready to ask them as my knowledge/understanding of MDT is still limited).

For clarification:

• through MDT you get boot.iso image with WinPE which holds all updates
• you boot VM with this image (in my case WMware Workstation)
• booted image has an idea how to connect MDT server and access ‘distribution’ folder with rest of the files
• process continues as if it was run from local media (DVD)
• you push drivers out (please clarify how this process works)

Ok, somehow I missed that MDT shares a folder with distribution files and generated WinPE (iso file) has some script which accesses (mounts) shared folder with distribution files make them available on installed computer and continue installation process.

Please confirm that my understanding is correct.

Then I try understand why in articles I read they mentioned WDS.

On side note: a few years ago I learnt about iPXE and tried diskless installation of Windows 7 which was successful (diskless Windows 10 installation did not work at that time – at least I could not find any reference).

Thank you again,
Polar Bear

@george1421
George, thank you for quick reply. I still try to find ‘head and tails’ if my process of creating gold image correct.

Just for clarification:

• where do you install MDT - is it reference computer or some other computer.

Through the reading I mentioned earlier they use WDS (Windows Server is required) instead FOG.

In our case we need to prepare reference computer with all updates installed and ready to be sysprep applied.

I do take an image at this stage before sysprep run (just in case if I have to return to this point).

Then I run sysprep with a shutdown and set a task to capture sysprep image (which can be installed on any platform – but really as you mentioned we need two images BIOS and EFI).

Prior this moment I used ‘Windows System Image Manager’ to create autounattend.xml file for both BIOS and EFI partition layout (again here we have x86 and amd64). In Windows System Image manager I specify that at early stage WinPE wipes hard drive and recreates partitions (for BIOS and UEFI) and then specify what partition to install the image into.

Now as we capture syspreped image with FOG and then deploy it into real computer I have obscure understanding what really happening at early stage WinPE (1).

As I see from information on the screen, FOG deployment restores two partitions for BIOS and I believe we already passed stage where installation recreates partitions (otherwise installation would wipe the disk with deployed image).

Now I try to understand how you use MDT to deploy image on reference computer. At the moment I only see it possible if MDT installed into reference computer itself. And all those updates will unnecessary inflate sysprep image.

In this case when we capture the image we will have MDT installed in each computer? Can you clarify this part of the process?

I do understand that WSUS offline repository can be stored on shared network disk/folder and it will not get captured into reference image.

I try to understand how to refine the whole process of preparing, capture and deployment of images.

Thank you again,
Polar Bear

@george1421

Thank you George,

I did some reading on MDT, WDS and WSUS…

Deploy Windows 10 with MDT 2013 and WDS
https://theitbros.com/deploy-windows-10-on-windows-server-2012-r2/

Capture Windows 10 Image using MDT 2013
https://theitbros.com/capture-windows-10-reference-image/

Integrate WSUS Offline Updater with MDT 2013 to Deploy Windows 10
https://theitbros.com/integrate-wsus-updater-with-mdt-2013-to-deploy-windows-10/

I understand that MDT can ‘embed’ upgrades into wim file for installation.

At some paragraph in articles mentioned above they mentioned ISO files and it is not clear if updates can get slipstreamed into ISO file which latter will be used for installation.

Do you have any information using WSUS to download updates which then copied to MDT to slipstream them into ISO image? This section was not completely ‘transparent’ if such procedure possible.

In the past I was slipstreaming SP3 into Windows XP, then using additional software to slipstream updates (but unfortunately not all of them, some left behind) and on top of it using slipstreaming (really embedding) drivers for most hardware.

Well, yes it still was time consuming process to build installation DVD but all this time was recovered during installation (at time I was employed with University and on positive side I embedded installation key so that graduate students could install Windows without IT group help on their lab computers).

Otherwise to install a computer with all required software would require a whole day – not best way to use work time.

George, I would appreciate if you could share in very short form what best works for you with mentioning your ‘hardware base’ (how many computers, how different computers).

For example, yesterday I tested Windows 8.1 gold image on Dell G31 mainboard with GMA3100 video card which is not supported. I have downloaded driver from Intel for G31 video, but it refused to copy files. I have downloaded video driver for Dell Vostro 530 and attempted to install it – again installation failed. And as last resort I did video driver update through Device Manager by pointing to Dell’s driver directory. After about a minute Device Manager somehow picked up the video driver (although driver is for Windows 7).

How do you handle drivers installation for old equipment which is not supported in Windows release?

In my case I use IObit Driver Booster on reference computer to install it into the image, and if the reference computer is not virtual machine then through it I update drivers (Driver Booster does pretty good job).

The deployment of such gold image on same hardware as rule does not require any update for drivers and in most cases the only update required is Defender or Essential security update (I got an impression that these updates sometimes released a few times a day).

Thank you for all your input,
Polar Bear

george1421,

How well works MDT?

I never tried it before and I do not mind to give it ‘a shot’ if it works well.

Thank you,
Polar Bear

Hi george1421,

what did you use to slipstream updates into wim file? How much time did it take?

I found that it is much easier to update reference computer with Windows 7 than with Windows 8 (I spent a few days to find a way to update Windows 8 – without WSUS it just did not work).

Yesterday I did a run test with Windows 8.1 – I have reimaged with gold image a real computer. Then I found that it detected a few updates – but attempt to install them failed (through windows updates check).

Thank you,
Polar Bear

Hello,

currently I work on fog deployment of Windows XP/7/8.1/10 and came across a problem with updates for Windows 8.1.

Details:
I prepare reference computer for sysprep on Windows 8.1. The system installed Ok, but windows update process does not work properly – the system looks for updates and never finishes the process of identifying required updates.

I’ve attempted a few potential solutions ‘found on internet’ for this problem and none of them produce desired result.

At this moment I could not find anything better than to use WSUS for offline updates – instruction.

Is there more ‘elegant’ solution to this problem?

I would like to keep an image of system in the state before running sysprep, so that I could easily install updates and then run sysprep to get ‘golden’ image with autounattend.xml.
(autounattend.xml file added to sysprep as an option works properly and been tested and in VM as in real computer)

Thank you for your help,
Polar Bear

@george1421
George,

hmm, something interesting have happened – I have reinstalled Debian and Fog once more.

And this time when I was creating new image in web interface I found that now instead of dashes ‘-’ replacing spaces ’ ’ it has removed them!!! Does anybody working on code right now? I thought that if release 1.5.7 came out then it will not be changing (except patches to fix faulty pieces of code).

I have created one image and then almost immediately attempted to create another one and went to grocery store. On my return back found that that second image was created fine and now in imaging logs it listed with dates.

For me it is a mystery, on my part I have repeated same steps without any deviation and if earlier it produced described error now it works fine.

I will run more tests to see how reliable current setup.

Polar Bear

@george1421
George,

this test computer is more than 10 years old when Windows XP still was a major Microsoft OS

https://support.hp.com/ca-en/document/c01680017

an article for a date reference

https://www.pcworld.com/article/132004/article.html

I found this computer disposed and decided to give it a second life, yesterday it misbehaved - I looked at mainboard and found two bulged capacitors – nothing what solder iron can not fix, few minutes and little termo-paste and computer runs again.

Half an hour to capture image is not too bad…

Polar Bear

@george1421
George,

just for a test I decided to capture image with compression index 6 (default) and average speed was 1GB/minute (sometimes hit 1.2GB/min).

If before to capture image with compression index 8 it was taking 51 minutes, now with compression index 6 it takes close to 28 minutes.

Polar Bear

@george1421
George,

I did install Fog with all defaults and I believe that https was not part of it (otherwise I did not touch https myself – may be in the future but not at test stage).

At this moment I use defaults in Fog web interface, no special storage groups was created – I use default storage.

I’ve thought about may be try to create new storage group and try to use it instead of ‘default’ – but I believe that ‘default’ should work anyway, otherwise this problem should be filed as a bug.

Polar Bear

@george1421

Well George, I do not use special software to measure performance of the HDD – for me numbers are coming from ‘hdparm -tT /dev/sd?’ tests on 775 socket.

Although I prefer WD drives in this case I got for good price Seagate 2TB SATA drive (Seagates was not so reliable in comparison with WD in my practical life).

Just for notice: I run Linux server with 4 WD Blue HDDs 2TB each in RAID 5 configuration and performance on read side is around 380MB/sec.

Polar Bear

@george1421
George,

well I installed Debian 10 essentials+ssh-server+sudo only – no firewall, no selinux.

I guess that problem is somewhere else. So far I did not see obvious reason which could cause described by me problem. What is most interesting that for first image it works fine, but when I try to create second image it fails… puzzling

Would you suggest to dive into ‘debug’ mode?

NOTE: for a test I attempted to install Fog 1.5.0 just to see if it does not ‘manifests’ this problem and found that installation has failed (last message does not clarify the reason – I did not looked into installation files).

Right now ‘Fog server’ under automatic re-installation of Debian 10 … (takes about 15 minutes to get it running again).

Polar Bear

@george1421

In my old computer I have 160GB HDD which should give read speed around 120MB/sec.

For a second we will not take into account CPU compressing data and we should come to a number of 7.2GB/minute.

Now let’s drop AMD Athlon X2 3800+ with performance index of 999 by Passmark and compression kills capture speed to a range of 580-650MB/min. (Note: I have chosen compression index of 8 instead default 6 for no obvious reason)

https://www.cpubenchmark.net/cpu.php?cpu=AMD+Athlon+64+X2+Dual+Core+3800%2B&id=77

This is my observation so far.

Polar Bear

Hi blong10,

I am in no way related to Fog project but see below my observation.

I currently testing Fog between two computers with 1Gbit network cards connected through 1Gbit switch.

Fog server captures Windows 10 Pro image, HDD has two partitions.

Fog identifies first partition (about 500MB) and make decision to not compress it, with second partition it compresses data (on client side if I understand it correctly) and sends over network. I observe sustained speed about 590MB/min and to capture Windows 10 Pro on 160 GB HDD takes about 51 minutes (CPU AMD Athlon X2 3800++ – low performance by modern standard).

When I deploy this image back to the Windows 10 machine it takes about 6 minutes 45 seconds.

My vision is that due required compression CPU of the host had to work hard and network speed at this time is no concern.

I would assume that in your case probably some data read into the buffer then they compressed (during this time you do not see network activity) and only then spilled into network wire.

Probably more knowledgeable people in this forum will give more detailed explanation of your observation.

Polar Bear

Hi,

next attempt failed as well, please see included pictures

Polar Bear

Hi george1421,

thank you for quick return to my question

Just a quick update:

1. I have woken up my computers (3 linux boxes: DHCP/DNS servers, webserver/storage/tftp, fog server (storage/tftp/apache for fog).
2. found that fog server did not waken up properly (something related to nvidia video card) and I had to reboot this computer
3. in fog set new task to restore original image (image without fog client)
4. to my surprise found that Windows 10 host rebooted (it had fog client installed at that moment) – odd, it did not work before I rebooted fog server
5. image deployment went without any surprises well, no errors and took slightly less than 7 minutes (computer quite old by modern standards)

At this moment I did not reached yet point of ‘golden image’ prepared with sysprep – I am testing fog, how it works to be assured that it will not bring surprises at production deployment.

I surely will return to sysprep step as only I get assured that I do understand fog procedures properly and no quirks will come up any more.

I run capture of Windows 10 image by rebooting Windows 10 computer. I believe that at reboot there is no ‘unclean’ filesystem and fog removes swap and hybernate file – at least this is what I see on the screen of host machine which goes through fog capture process.

I will verify that there is no ’ -’ combination when I create the image.

I open web form through web browser and type ‘Win10 Pro 1803 upd Aug 1, 2019 client’ and field in image name filled automatically for me (I believe javascript takes care of this part and I believe that fog developers have looked into the code to handle situation for invalid characters – at least I confirm that I did not typed image name manually).

I did not spent sufficient time for ‘think process’ of how to name images – what came to my mind to have the name descriptive enough to distinguish what type of content it has. Any idea on this subject is welcome – I am very flexible in adopting ‘good practice’ approach.

Right now I plan to make a new attempt – who knows may be fog server restart has ‘kicked in’ something what did not worked earlier.

I will report as soon as I will get new results.

Thank you again
Polar Bear

Hello,

I am experiencing some problem with latest Fog 1.5.7 installed in Debian 10 (‘Buster’).

The installation went through smoothly, Fog server is up and running. I can register host(s) and I can capture ‘first’ image from Windows 10 machine (HP dc5750 SFF). The image can be deployed back to the machine. Windows 10 image (win10pro-upd-aug1-2019) is clean install with all updates on Aug 1, 2019.

Now I wanted to install Fog client and to make a new image. Fog client was installed in Windows 10. Fog server was used to create a new image (win10pro-1803-upd-aug1-2019-client) and host scheduled to capture new image immediately.

Host with Windows 10 had no user logged in, but scheduled capture did not started (Fog client did not reboot the computer). I’ve attempted ‘force to start task’ and nothing has happened – I waited for a while to be assured that reboot will not take place.

Then I rebooted host with Windows 10 myself, Fog began to capture new image and everything was fine until partclone was working. Then Fog attempted to update database and it failed, it attempted to reapply changes and it failed again and again. Then Fog informed that host will be rebooted in 1 minute… and on next boot Fog initiated capture process again… it becomes endless loop.

Last messages on the screen indicate ‘invalid Storage group’ but it initiates on ‘Database update’ which fails.

I am sure that partclone did it’s job perfectly, but for some reason Fog fails to communicate/update database records. In web panel I can see a size of the captured image but date/time is wrong – if to believe that information it took 2019 years 8 month couple of days and a few hours to capture the image.

During Fog installation all options are default (I use other DHCP server and answered No on question if I would like Fog server to be DHCP server).

Some an idea came to my mind that may be there is some limitations on image name, I’ve decided to run a test with other image name ‘win10-test’ and the outcome was the same.

I did several attempts, even I have reinstalled Debian 10 and reinstalled Fog just for an assurance but result is the same – first image captured fine, but second fails with same message.

Did somebody came across this situation and is there some remedy to it?

Is my understanding is correct that when a new task (image capture/deployment) is scheduled ‘immediately’ through Fog client the computer (host) should be restarted automatically and I do not need to ‘force the task’ through web interface?

https://wiki.fogproject.org/wiki/index.php/FOG_Client

— Quote --------------------------------------------------------------
Task Reboot – This will just check if the client is in a tasking (other than a snapin tasking). If it is in a tasking, and the module is enabled, the host will be told to reboot. There is a third portion though in that if the user is logged in, and enforce is not enabled nothing will happen.
— End of quote ---------------------------------------------------

Thank you in advance
Polar Bear