Failed to Set Disk GUID (sgdisk -U) (restoreUUIDinformation)

[Sebastian Roth]

@Junkhacker said in Failed to Set Disk GUID (sgdisk -U) (restoreUUIDinformation):

edit3: ok, that was giving me kernel panics. not sure what was going on there

Did you get this fixed??!?!

[Junkhacker]

@Sebastian-Roth only by reverting back to previous inits

[Sebastian Roth]

@Junkhacker Please go to the FOG web UI -> FOG Configuration -> FOG Settings -> FTP Server and increase KERNEL RAMDISK SIZE from 127000 to 275000.

Sorry for all the rumor.

@Tom-Elliott Not a good situation where we would need to either revert the inits to 100 MB size or quickly push out a new release just to fix the KERNEL RAMDISK SIZE. I will build 101 MB inits (yes, just a little larger to hopefully get around the space issue mentioned below) over night just so we have those handy.

Edit: Good news, we still had the tool chain from the latest jenkins build on the server and so I am able to generate new inits fairly quickly. Will update the latest ones on the web server as well as binaries.1.5.6.zip for now.

Edit2: inits and binaries1.5.6.zip are updated. New checksums are:

e69e0763cfda1e02159553ea1cdbdc55  init.xz
00d811eef21cdfab4bab5160ad888bec  init_32.xz

[Tom Elliott]

@Sebastian-Roth well if the workaround is to update the ram disk size, it’s not too hard a thing to do. We can push out a patch to fix it directly, but seeing as this value is already easy to change what about simply documenting the error and the change for the resolution?

As fog keeps making progress I don’t think asking people to make this change is too difficult. And as for 1.5.7, just add in a patch to check if the ram disk size is too small in the bootmenu, and if so update the value directly.

[Sebastian Roth]

@Tom-Elliott Sure we’ll fix that in 1.5.7 but I’d like to see 1.5.6 not going on ground just a few days after we released it. Therefor I am trying to get to a solution where people can just use 1.5.6 without having to fix anything manually. Just tested and seems we are fine with the currently uploaded files. All good I hope.

[austinjt01]

@Sebastian-Roth So I re-ran the FOG installer. After that I tried capturing an image which appeared to work okay, but now i’m getting the same error upon trying to deploy that captured image.

It goes through the whole download process and says “A warning has been detected: Failed to set disk guid (sgdisk -U) (restoreUUIDInformation)”

Sits on that screen for about 60 secs and then gets stuck at “Changing hostname…” again

Shall I try manually downloading the files ? Maybe the re run of the FOG installer didn’t cut it?

Thanks.

[Sebastian Roth]

@austinjt01 Please run the following commands on your FOG server and post output here:

md5sum /var/www/html/fog/service/ipxe/init*
cat /images/<IMAGENAME>/d1.partitions

Make sure you put in the name of your image instead of <IMAGENAME>!

[austinjt01]

@Sebastian-Roth

I have our images stored on our media server, if I path over to /mnt/media-images/<IMAGENAME>/d1.partitions

This is what’s in the file

UPDATE:
I manually installed the updates and got this as the output

[Sebastian Roth]

@austinjt01 said in Failed to Set Disk GUID (sgdisk -U) (restoreUUIDinformation):

So I re-ran the FOG installer.

The md5 sums we see in the picture are definitely not the most current ones. Which installer version did you run? Where did you download FOG? I mean from github or downloaded an archive?

[austinjt01]

@Sebastian-Roth I ran the updates and got the same output that you have below.

Seem to still have trouble with capture / deploy though. Keeps getting hung up on “Changing hostname…”

I manually installed the updates and got this as the output

[Sebastian Roth]

@austinjt01 said in Failed to Set Disk GUID (sgdisk -U) (restoreUUIDinformation):

Seem to still have trouble with capture / deploy though. Keeps getting hung up on “Changing hostname…”

What do you mean by “hung up”?? Please take a picture of the screen and post here.

[austinjt01]

@Sebastian-Roth

It will just sit here and not do anything.

[Sebastian Roth]

@austinjt01 This is the exact same picture you posted days ago. I can hardly believe you see the exact same screen again as we have changed the size of the init files and I would expect the “No space left in device” error to have vanished.

Please take a new picture of what exactly you have on screen when it hangs!

[austinjt01]

@Sebastian-Roth Yeah you are right, I grabbed a different laptop to test on just now and everything is fine. Thanks for your help, I really appreciate it and your support.

Edit: The capture worked this time, but the deploy gives this error.

Picture was taken just now

[Sebastian Roth]

@george1421 @Junkhacker Is anyone able to to reproduce this issue?!?!?

[george1421]

@Sebastian-Roth said in Failed to Set Disk GUID (sgdisk -U) (restoreUUIDinformation):

@george1421 @Junkhacker Is anyone able to to reproduce this issue?!?!?

I can not duplicate the issue here, but that just means the circumstances (environment) is different.

I did see something that raised a red flag.

I have our images stored on our media server, if I path over to /mnt/media-images/<IMAGENAME>/d1.partitions

How are you doing this? You are not supposed to be able to reshare an nfs mounted share (technically you can, but fog is not configured to support it). You can’t do that in ms windows either.

Images are captured to /images/dev/<mac_address> directory (nfs share) and then “moved” to the /images/<image_name> directory using the ftp server built into the fog server. At the very least if /images/dev is on the fog server and /mnt/media-images/<image_name> is on the NFS share the ftp server will have to copy the entire image to the remote nfs server instead of just moving the file pointers as it would if the /images/dev and /images are on the same disk. If you want to store your images on an NFS nas but have the fog server manage the process there is another (unsupported) configuration you can use instead of resharing an nfs mount.

[austinjt01]

@george1421 Our storage node that we have setup is on ‘media’
‘/mnt/media-images’ is just an admin shortcut for us to navigate easier we aren’t using that for uploading or downloading. Upon capture the storage location on the FOG screen is listed as: ip address:/mnt/pool/it/images/dev/ Is it a possibility that the image is being captured into DEV but not being copied out of dev during deploy?

The other thing I have to add to this, I have 7 different brands of laptops. All of the other brands work (Capture and Deploy) just fine. These new laptops we just got (30, Lenovo P52s) they are at least two years newer than any other device in our inventory. Do you think they need to be set up on UEFI instead of Legacy boot?? The only drastic difference between these machines is that they all have SSD’s which I didn’t think would make a difference?

Thanks

[george1421]

@austinjt01 is 10.228.255.10 your fog server?

[austinjt01]

@george1421 no, that’s the IP of our file server. .3 is the FOG server IP

[george1421]

@austinjt01 So your file server is a linux server with the fog code installed? Or is it a linux (like) nas device?

Do you have this working using the non-supported configuration where you create a second storage node and then set that node as the master? I’m still trying to understand your landscape.