Fresh install of 1.2.0 on Ubuntu 15.04 - Imaging suddenly "broken"

[theterminator93]

I just finished setting up a fresh install of FOG 1.2.0 on a fresh install of Ubuntu 1.2.0 this morning. I began setting up image definitions in the console, then registered a few hosts to upload images from. I started by uploading one image, a Dell Optiplex 380; completed fine. I started uploading a second image, an old IBM ThinkCentre S50, completed fine. While I was waiting for software to update/optimizations to complete prior to uploading additional images, I successfully deployed the just-uploaded ThinkCentre image. Then I uploaded an image for an HP DC7700. While this image was uploading, we tried imaging more ThinkCentres; suddenly they were not imaging properly.

The host registration completed fine, it PXE booted fine, and all the pre-partclone FOG preparations finished, up to the point where it would erase the partition tables, restore the MBR, check hard disks etc. At this point it would normally start PartClone to start pulling the image down. Nothing; a few seconds passes and then it suddenly says it’s changed the hostname, updating computer database status, imaging complete - despite not actually imaging anything. I’m left with a blank hard disk.

I tried imaging a couple Dell 380s using the images I had uploaded, same thing. All of a sudden the image definitions in FOG say the image size is 0.00 GB on the client and 2.01 GB on the server.

I’ve done an in-place reinstall of FOG, no luck. I’ve done chmod 777 /images, no luck.

What’s going on? What caused this thing to break after imaging one computer?

[Wayne Workman]

What version of windows are you imaging? and is the disk MBR or GPT ?

[theterminator93]

All Windows 7; standard Windows 7 partition setup with the hidden system partition and main data partition. FOG set to multi partition, single disk.

I read a few things about potentially corrupted images, but that still doesn’t explain (to me) why imaging broke after successfully deploying an image to a client; would it be worthwhile if I downgraded to 14.04? My other FOG servers are running this OS without issues…

[Wayne Workman]

@theterminator93 Well, you’re lucky because there is known instability with using 1.2.0 on Ubuntu 14.04 and 15.04 as well.

Are you positive that a size is showing for size on client and not for size on server??

Can you post the output of this command ran on this particular FOG box - it’ll be a lot of output, use a code box:

ls -lahRt /images

[theterminator93]

Here is the output of the command…

/images:
total 32K
drwxr-xr-x 25 root root 4.0K Jul 23 14:39 ..
drwxrwxrwx  8 root root 4.0K Jul 23 14:31 .
drwxrwxrwx  2 root root 4.0K Jul 23 14:31 dev
drwxrwxrwx  2 root root 4.0K Jul 23 13:50 StudentOpti380
drwxrwxrwx  2 root root 4.0K Jul 23 13:09 StudentDC7700
drwxrwxrwx  2 root root 4.0K Jul 23 12:14 StudentIBM
drwxrwxrwx  2 root root 4.0K Jul 23 11:23 LabOpti380
drwxrwxrwx  2 root root 4.0K Jul 23 10:36 postdownloadscripts
-rwxrwxrwx  1 root root    0 Jul 23 10:36 .mntcheck

/images/dev:
total 8.0K
drwxrwxrwx 2 root root 4.0K Jul 23 14:31 .
drwxrwxrwx 8 root root 4.0K Jul 23 14:31 ..
-rwxrwxrwx 1 root root    0 Jul 23 10:36 .mntcheck

/images/StudentOpti380:
total 12G
drwxrwxrwx 8 root root 4.0K Jul 23 14:31 ..
-rwxrwxrwx 1 root root  12G Jul 23 14:31 d1p2.img
drwxrwxrwx 2 root root 4.0K Jul 23 13:50 .
-rwxrwxrwx 1 root root 8.3M Jul 23 13:50 d1p1.img
-rwxrwxrwx 1 root root  512 Jul 23 13:50 d1.mbr

/images/StudentDC7700:
total 13G
drwxrwxrwx 8 root root 4.0K Jul 23 14:31 ..
-rwxrwxrwx 1 root root  13G Jul 23 14:13 d1p2.img
drwxrwxrwx 2 root root 4.0K Jul 23 13:09 .
-rwxrwxrwx 1 root root 8.2M Jul 23 13:09 d1p1.img
-rwxrwxrwx 1 root root  512 Jul 23 13:09 d1.mbr

/images/StudentIBM:
total 7.3G
drwxrwxrwx 8 root root 4.0K Jul 23 14:31 ..
-rwxrwxrwx 1 root root 7.3G Jul 23 13:29 d1p2.img
drwxrwxrwx 2 root root 4.0K Jul 23 12:14 .
-rwxrwxrwx 1 root root 8.2M Jul 23 12:13 d1p1.img
-rwxrwxrwx 1 root root  512 Jul 23 12:13 d1.mbr

/images/LabOpti380:
total 11G
drwxrwxrwx 8 root root 4.0K Jul 23 14:31 ..
-rwxrwxrwx 1 root root  11G Jul 23 11:58 d1p2.img
drwxrwxrwx 2 root root 4.0K Jul 23 11:23 .
-rwxrwxrwx 1 root root 8.3M Jul 23 11:23 d1p1.img
-rwxrwxrwx 1 root root  512 Jul 23 11:23 d1.mbr

/images/postdownloadscripts:
total 12K
drwxrwxrwx 8 root root 4.0K Jul 23 14:31 ..
drwxrwxrwx 2 root root 4.0K Jul 23 10:36 .
-rwxrwxrwx 1 root root  233 Jul 23 10:36 fog.postdownload

The interesting this is immediately after I uploaded the images, the image page in the console told me the hard disk size required by the image in the image size “on client” field. The image size “on server” is always reported as 2.01 GB for some reason. The images I’ve tried to deploy now say the image size on the client is 0 GB…

Here is what the image management area looks like…

Image Management
All Images
Lab-Opti380
default:Windows 7
Multiple Partition Image - Single Disk (Not Resizable) 0.00 iB 2.01 GiB Yesterday, 3:58 pm Staff-Opti380
default:Windows 7
Multiple Partition Image - Single Disk (Not Resizable) 0.00 iB 0.00 iB 01-01-1970 12:00am Staff-Opti980
default:Windows 7
Multiple Partition Image - Single Disk (Not Resizable) 0.00 iB 0.00 iB 01-01-1970 12:00am Student-DC7700
default:Windows 7
Multiple Partition Image - Single Disk (Not Resizable) 74.33 GiB 2.01 GiB Yesterday, 6:13 pm Student-IBM
default:Windows 7
Multiple Partition Image - Single Disk (Not Resizable) 0.00 iB 2.01 GiB Yesterday, 5:29 pm Student-Opti380
default:Windows 7
Multiple Partition Image - Single Disk (Not Resizable) 0.00 iB 2.01 GiB Yesterday, 6:31 pm Student-Opti755
default:Windows 7
Multiple Partition Image - Single Disk (Not Resizable) 0.00 iB 0.00 iB 01-01-1970 12:00am ```

[theterminator93]

Additional development.

I flipped our DHCP settings to use one of the other “tried and true” FOG servers we’ve been using for a bit over a year, and I uploaded a couple images to it. As soon as we tried to deploy one of the images (the IBM image), the workstation finishes the PXE init, loads the FOG kernel and starts to image normally, gets to the point where it tries to start PartClone to download it, then the image file appears to go corrupt (e.g. image size-on client value goes from the HDD size to 0 GB) on the server and everything fails. I’m thinking less a server issue and more a client or hardware issue now, but why would it have successfully deployed once, then failed every other time? And why is it corrupting the image on the server (at least in the regard of wiping out the deployed size definition)?

[Wayne Workman]

@theterminator93 I actually have two Optiplex 380s at my home (I have a fog server at home and at work) and I can confidently say that FOG Trunk can image this model with an MBR setup using Win7 without an issue.

Given that you’re reliable FOG server failed to image this model, I’d recommend that you move to FOG Trunk and then retry with this model. Upload and then download.

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

Please use the Git method, as SVN is currently unavailable.

You may update an existing FOG server - or build one from scratch and use the Git method when it comes time to put FOG on it.

[theterminator93]

I’ll try the new trunk release and see what happens on Monday. I actually replaced the HDD in the PC that was running in the FOG server at this site and was in the process of uploading updated images from the various PCs there. They had all been imaged without issue on the old server, which was also running 1.2.0 (although it was Ubuntu 11). Strangely though, we’ve never had any issues in the last 3 years with the various releases of FOG imaging the PCs in this setting (I estimate the number of images pushed down to be well in the 2000+ range), all with the same hardware. Especially considering I got one successful deploy on a machine before it stopped working on all hardware. That’s why I’m so baffled as to why it stopped working all of a sudden.

[theterminator93]

No luck. Re-uploaded the image on a fresh install of Ubuntu 12.04 with FOG 1.2.0 and attempted to deploy - same problem. Upgrade to FOG Trunk and attempted to deploy, same problem. An error message would be helpful but none are seen…

I’ve just about given up. I’ve had my techs start imaging these machines with Acronis so we don’t keep falling behind on our timetables. If anyone has any other ideas or things to try, I’m all ears.

[Wayne Workman]

@theterminator93 said:

If anyone has any other ideas or things to try, I’m all ears.

Might sound harsh but - ditch Ubuntu, install Fedora 21 or 22 Server (or CentOS) - Follow this guide: https://wiki.fogproject.org/wiki/index.php/Fedora_21_Server

And if you have any questions at all about Fedora or that guide, just message me.

[theterminator93]

I will give that a try later this week and report back. I’ve got nothing to lose at this point.