what is the usual cause of corruption in the images? over the last 6-8 months we have had a few go bad, I would just recreate them if it was a one off, but this seemed to be a bunch so I wanted to make sure it was corruption before reinstalling.
We are running identical hardware (both with the same redundancy/mirroring) at another school and have had zero issues.

They’re configured under FOG Configuration->FOG Settings. FOG_TFTP_PXE_KERNEL should be set to bzImage32 and FOG_PXE_BOOT_IMAGE should be set to init_32.xz

They’re stored in {fogwebdir}/service/ipxe/

Hope this helps.

Well, this isn’t necessarily the case. Depends on the layout.

If you create an image as multipart non resizable off of a fresh cd install of windows 7, the d1p1 and d1p2 files, for all purposes, are rec.img.000 and sys.img.000 respectively. So you could, potentially, switch this image from non-resizable to resizable. However, if you create a resizable image, you are kind of stuck. This is because the mbr is not available.

If you were to go from multipart to resizable, you’d need to rename the d1p1.img and d1p2.img on the server.
Rename d1p1.img to rec.img.000
Rename d1p2.img to sys.img.000

I hope this helps.

What?

If it passed the first time, but not the second, then something would possibly be wonky within the BIOS is my guess. There’s no changes made to the system or kernel during the registration phase.

and those problems were solved with…? replies without answers are worthless. (my statement isn’t in a mean tone)

The best way I found is to export the appliance, unfortunately. Unless you only want to make copies in which case cloning would be the best solution.

I use an external drive to store all my Virtual Images on so I don’t actually “export” them I just set up the appliance with the current files at my other locations. This allows me to have the same image and upload to the different servers (I chose stand alone servers instead of Nodes… not really sure why, I was just new at setting FOG up and I never re did the set up I know it’s going to change when I upgrade to 0.33 version but for now I just have multiple servers.)

I am partial to the snapshots feature. That and I use fairly antiquated technology, the ones that are on the phase out, for my imaging so I never had a GOOD server to build ESXi on, so the bad taste left in my mouth may be my own fault, but I never had good luck with VMWare.

Which Revision of Ubuntu?

I can confirm that 13.10 works fine. There are ALWAYS issues to over come when launching FOG in each environment, everyone has different switches and hardware. I know you want to blame Ubuntu, and I will admit that the latest revision of Ubuntu have needed some adjustment to work perfectly with fog. But I can also confirm that each one of my test servers is at the latest revision. So I don’t think this is a problem with Ubuntu.

I also want to note that I am able to boot, and image to my servers on 13.10, but I am using the Beta version of FOG.

There you go! That’s where you need to make your changes, with the Web GUI!

It works! Thanks so much for your help!!!

Thank you, Tom. The issue is solved. 🙂

Have you tried re-downloading the installer in case it’s corrupted? Also are you using version 32 or 33b for FOG?

Take a look at this post and troubleshoot a little on your end. It sounds like either the mysql server isn’t running, or the extension may not be installed.

[url]http://ubuntuforums.org/showthread.php?t=396974[/url]

Made sure that the node user/pass, tftp user/pass, and linux user/pass were the same. My Fog server is now back in business.

Thanks!

It has been a few weeks since I made the original post (time flies when new projects are added). The above posts pointed me to a viable solution (although I wouldn’t consider it ideal) by manually adjusting options 066 & 067 in our 2008 server handling DHCP for the clients booting to the LTSP server. By default our DHCP server points clients to the Fog server. The DHCP settings for the LTSP clients have been manually added to point to the LTSP server.
I would have preferred a better level of automation, but for the time this solution works. Thank you to Vincent & Jaymes for taking the time to help.

Another option is to use one my kernels. They don’t have modeset enabled and should not cause this issue anymore:

[url]https://mastacontrola.com/fogboot/kernel/bzImage[/url] (64bit) or [url]https://mastacontrola.com/fogboot/kernel/bzImage32[/url] (32 bit)

As you’re getting the error when starting the task, what do you get when you try to create a simple directory in /tftpboot/pxelinux.cfg/?

My guess is the permissions are not correct.

To make sure all is fine in the tftpboot directory make sure you run the command:
[code]chown -R fog:root /tftpboot[/code]

This is a very simple command that I find people saying they’ve performed, but in actuality, the they’ve only run:
[code]chown fog:root /tftpboot[/code]

While they look the same, I assure you they are not. The -R tells the chown command so set everything within the directory to change ownership as well.

[quote=“Tom Elliott, post: 23986, member: 7271”]Do me a favor and remove the cd from the drive and try again.

Something tells me the image it captured is from the drive itself, rather than your actual hard drive, though I don’t know why it would. Can you try using my kernel as I don’t allow it to read the cd drive at all, this should mean you don’t have to worry about sr0 trying to load as a viable drive.[/quote]

Also, as I see you are a developer, all of the PCs I am going to deploy to (as well as the reference machine) are UEFI… Will this cause an issue with a partion image or should I DD them to make sure UEFI data remains untouched? They are all Compaq 6200 pro boxes and I have found there is no way to set PXE as a default boot option when EFI is in play, they will not boot to windows wihout the EFI windows boot manager source being selected and network is only in the legacy section which cannot be prefered over EFI source :-[

.32 doesn’t like to back fill information, if you made the host BEFORE you entered the AD information under fog settings you will need to copy paste it into the hosts. Any hosts you create going forward will have this information filled in.

You could also delete the hosts and rejoin them to FOG.

Make sure if you do this to select “y” for “Would you like to join this host to AD using default info?”

Have you tried running a quick inventory on the 9 hosts that did not show up? Running a quick inventory or doing a “Full host integration” will load the host into inventory. Quick joins will not.

This isn’t a kernel issue at all, as the Services (GreenFOG, TaskReboot, etc…) don’t do anything with the kernel/linux layer of the FOG system.

The only one that, might possibly, have anything to do with the kernel is the Task scheduling shutdown, but even that I’d be hard pressed to believe.

I think this issue is more related to the way the shutdown command for the task cycle is read. I don’t know much about 0.32’s init.gz file but it may be trying to see shutdown as either 0 or 1, but the shutdown sent when creating a task is set as shutdown=on. I could be wrong and will take a look for you.