Image deploy error - No Drive Number Passed



  • fog_issue.jpg ![0_1564049049220_fog_issue.jpg](Uploading 98%)

    Hi

    Can anyone advise on this error.

    The image is a multi disk/multi partition image

    I’m running fog 1.5.7 on centOS 7

    I have captured the image but getting this error when attempting to deploy.

    Any help much appreciated.

    thanks


  • Senior Developer

    @psmith10 said in Image deploy error - No Drive Number Passed:

    but the Ubuntu image (on the SSD) was not deployed.

    Did you see an error message or something else in the output scrolling past that would give us a hint on why? As well you might do a debug deploy and check the messages there.



  • @Sebastian-Roth I removed the .size files from the image directory. The host then accepted the image without the error but the Ubuntu image (on the SSD) was not deployed.



  • Thanks for the reply. I am currently running off the 1.5.4 fog server but when i get a chance i’ll spin up the 1.5.7 server and try deploying the image with the .size files removed.
    Regards
    Paul


  • Senior Developer

    @psmith10 Some months ago we added a “fix” for someone having two NVMe drives in their PCs and Linux enumerating those two drives in totally random order. The fix was to also capture disk size information and try to match disk size on deploy to push the right image to the correct disk on deploy. So what happens in your situation is that it cannot match the disk size on systems with the slightly smaller Western Digital drives and fails.

    See the details here: https://forums.fogproject.org/topic/12959/dell-7730-precision-laptop-deploy-gpt-error-message

    Good you brought up this topic because I just stumbled upon another topic the OP brought up that seems we have lost track of this. We may need to re-think the whole mechanism.

    On the other hand there quick and dirty fix for your situation where one drive is NVMe but the other one is a SSD. Simply remove (move to a different directory) the two files named d1.size and d2.size from your image directory (/images/<IMAGENAME>/) on your FOG server. Then try deploying again. The scripts should do normal disk counting when the size files don’t exist.

    Windows reports that the Integral drive is 111.79GB and the Western Digital drive is 111.80GB, so there is a discrepancy.

    Sounds like you are very lucky! Would it be the other way round you wouldn’t be able to deploy in “All Disks” mode even if I’d fix the issue mentioned above. If the destination drive is only one sector smaller than the source it won’t be able to deploy in non-resizable mode (resizable All Disks has not been implemented).



  • @george1421 The sda drive is a 120GB SSD.

    I have double checked and the image definition is the same on both old and new fog servers.

    The source system has an Integral 120GB SSD

    Following further tests, target systems with the same Integral 120GB SSD accept the image without an error.

    I have a batch of systems with a Western Digital 120GB SSD. These are all throwing the error.

    Windows reports that the Integral drive is 111.79GB and the Western Digital drive is 111.80GB, so there is a discrepancy.

    The image is captured from the smaller Integral drive.

    Another thing to note is that Windows recognises the Integral drive as Disk 1 and the WD drive as Disk 0.

    All target systems are the same hardware and the two SSD variants are plugged into the same SATA ports.


  • Moderator

    @psmith10 I see that your target system has both a rotating hard drive (sda) and an nvme disk. Does your source system have the same type of disk layout?

    I find it interesting that its saying no drive number passed if the systems are identical. I may need to look in the code or get a developer involved because this is a unique error.

    Are you sure you have the image definition and host definition identical between the 1.5.4 and 1.5.7 installs of FOG?



  • As a test, i have spun up an old fog server running 1.5.4 on Ubuntu 16.04.

    I captured the same image from the same machine and deployed with no errors.

    Unfortunately the old server doesn’t have enough space, hence the upgrade, so any help regarding a fix would be most appreciated.

    thanks



  • fog_issue01.jpg
    Thanks for the quick response, this is the result of the command


  • Moderator

    Will you do this. Reschedule a deploy to this target computer. Before you hit the submit task button tick the debug checkbox. Now submit the deploy task. PXE boot the target computer. After a few screens of text on the target computer you will be dropped to the FOS Linux command prompt.

    At the fos linux command prompt key in the following command lsblk and post the results here. This will show the basic geometry of the target hard drive.


Log in to reply
 

322
Online

7.4k
Users

14.5k
Topics

136.7k
Posts