I’m sorry that I’m not right on the case.

I wonder about this issue. The specifics of the problem you’re running into seems, to me, to indicate that there is no Storage groups with a master node within the group. Normally it should default to return the relative node that IS in that particular group, but if there is no node associated, this could occur, I suppose. I don’t know all the specifics. I’m working, hopefully, to figure out a solution, but I feel I may need access to a problematic system so I can work on it live. Sometimes, and there are multiples who can verify this as needed, I simply can’t replicate so seeing the problem on a known “bad” target is really the best option.

Thank you.

OK, I will try this. Thanks!

@mrzo For the record - my organization has had a whole lot of problems with the Lenovo M72e. And when I say problems - they are (currently) absolutely not working with FOG Trunk.

The strange thing is - the 30 or so that we tried every method in the book with - one of them always succeeded while every single other one always failed… we have sat down and compared Firmware settings, we’ve restored firmware defaults… we can’t figure it out, it’s a mystery at the moment. We are currently using WDS and/or Norton Ghost for that particular lab.

We have no M83 models in our environment - but I wanted to just shed a little light on my experience with Lenovo. I think some of their other models don’t quite like San-Boot style HDD exits either… Lenovo is… a PITA.

AHCI is the preferred operating mode, it’s a best-case scenario that you have a mechanical drive.
You say that the system is configured to boot via UEFI ? What boot file are you using? Have you tried others?

@Wayne-Workman said:

Just to make it clear, Secure Boot is currently required to be OFF to network boot with FOG. Can you provide us with a screen shot of the error ? You might need to take a video to capture the error, often times the errors are only displayed for fractions of a second. What happens if you try to boot using ipxe.efi ??

At first, I didn’t know what you were talking about–the only efi file in my /tftpboot folder was snponly.efi. But then I read another suggestion from you elsewhere that maybe upgrading to trunk might help, so I tried that.

And, lo, my /tftpboot doth overflow with efi files! So I tried ipxe.efi, and it WORKS.

Also - see my old post below on the subject of using a USB dongle.

That looks like really useful info once I get this host registered. Amazing.

As you requested elsewhere, I will experiment with co-existing EFI and BIOS hosts using DD-WRT and see if I can’t contribute to the Wiki to pay it forward.

@hlcstudy Very happy you got it fixed and glad we could help.

@johnomaz said:

Yes, it can be flagged as solved.

Good 😉

@Wayne-Workman That is much more elegant than either of my solutions. I doubt you will need my help, but give me a poke if you want me to look at anything with you.

I’m going to solve this only because it seems a very particular set of information was causing the problem. While I was concerned my recent changes were the main cause, it seems they really weren’t, potentially.

I’m only solving because the problem was relatively easy to correct for.

Also - when did this failure start? Was it working before? What changed?

Have you checked the mysql settings inside of /opt/fog/.fogsettings on the problematic storage nodes?

Are you refering to Wake On Lan ?

I should explain a little perhaps - and then explain a few options.

When a computer is configured in it’s firmware to boot to the network first, It will try that. DHCP hears the request and sends info. The PC confirms it’s received info - there’s a fraction of a second’s wait - then DHCP sends a “Use it” command once no other computer reports “no don’t use that, I’m using that IP”.

Then the computer network boots to FOG - as it was told. If it is registered - it will look for any active tasks and do those - if there are no tasks it will either (A) Show the boot menu for a specified time and then try to boot to the HDD or (B) not show a menu at all and immediately boot to the HDD.

Just depends on how you have it configured.

By default, non-registered hosts will show the menu so that you see it’s not registered. This menu also has a countdown that elapses - after it elapses, the machine tries to boot to the HDD.

Now - if you don’t like seeing the menu for registered machines - the very best that can be done is to set the timeout to 1 second. You can do that here:
FOG Configuration -> iPXE Boot Menu -> Menu Timeout (in seconds):*

Now, do not set this to 0 because 0 makes the menu stay up forever… 1 is as low as you can go.

@Developers is that normal? 0 causing the menu to just stay up?

@middendorf said:

@ch3i

4557, like FOG says is the actual git version?

Can you rename the fog and html folders (like fog.bak and html.bak) and rerun the installer.

@Tom-Elliott absolutely, was looking for how I could mark resolved this side funnily enough,

I created a new installation of Windows without system reserved partition.

Apply the following steps when the Windows partitioning window apeares:

Create a partition on the blank hard drive. Select the partition drive as the partition to install Windows on. Click “Next”. The 100MB partition will be created. Delete the partition which you created in step 1. Extend the 100MB partition to the desired size. Select the partition which you extended in step 5 and select ‘Format’. After formatting the ‘system reserved partition’ will now become normal ‘system’ partition.

I can confirm this works 🙂

@ambers said:

We are currently using 1.2 on Ubuntu version 14.04

Maybe a problem with GPT, have you install from scratch the computer ? If yes have you remove the old partition table ?

@Wayne-Workman didnt testet it thoroughly, but as I described it, it didnt work, (I made sym link /image - so no change in exports was needed?) , permissions and files created were ok - after changed exports to new location, everything startet to work.

I wanted to respond with a link:

https://wiki.fogproject.org/wiki/index.php/BIOS_and_UEFI_Co-Existence

Please bear in mind that at the time of this post, this article is a work in progress - however at this point it does contain a lot of information and links to other information.

@josh.nicholson99 I should slow you down firstly - do your vlans go over slow and long distance links?

For example, my organization has every site connected via dedicated lines - but those lines are low speed links. We avoid heavy loads over the WAN links at all costs. EVEN THOUGH we have a vlan infrastructure that can easily bring any network to any port that we wish. If your organization’s setup is similar to this, I’d urge you to use the location plugin and setup storage nodes at each of your physical locations.

In the case that you have amazing WAN speeds or don’t have WAN links at all, then I’ve found these links that might be of interest:
https://communities.bmc.com/thread/41369
http://www.humboldt.co.uk/a-working-tftp-server-for-multi-homed-linux-systems/
http://www.tecmint.com/install-pxe-network-boot-server-in-centos-7/

It would seem as if dnsmasq is the answer… I haven’t looked deeply into it yet though.

Glad to see this is fixed. I helped to the extent of my abilities via forum chat but I came to a dead end basically. And Tom’s post explains why. Good to see a solution is found.

I love how this guy comes here - posts, gets impatient, posts again - then never comes back… and after he leaves, he gets 5 genuine pieces of advice from 5 different people - but he’ll never know that, will he?

All of us here have Jobs man! We are volunteers for the FOG Project!