• First, thanks for a great project. I am a newbie to FOG and have a lot to learn.

    I have not had any issue imaging any of our Windows workstations nor any linux machines on single hard drives.

    I have been reading about how to create an image for a linux machine that has a mdadm raid disk setup. As I understand the documentation I have read, setting the “kernel options” to “mdadm-true” in 'Host Management" is all that is required? Leave “Host Primary Disk” blank?

    This first client with mdadm is a simple two disk array (Raid1-mirror) linux machine. If I can start here, create an image, I can learn as I go.

    Sorry to ask this simple question but, most of the searches (“FOG Project mdadm”) keep finding “Intel Raid” or “hardware raid” or “fakeraid” references. Not “mdadm” specifically.

    Perhaps I overlooked the ‘mdadm client page’ in the wiki for this?

    Are my client settings the correct attributes to use?

  • Senior Developer

    @trebor said:

    Capturing the screen “lsblk” output from a pxe client, I am not sure how to do? Suggestions?

    Schedule a debug capture task for this client. Hit ENTER twice then you can run any command on the shell. Best if you can test mounting the RAID devices as well as getting a picture of lsblk from that shell so we see how FOG sees your system.

  • Thanks for you help on this.

    Yes, Ubuntu client.

    Mounting the raid devices (/dev/md126 & /dev/md127) I fairly comfortable I can figure out how to do that. Capturing the screen “lsblk” output from a pxe client, I am not sure how to do? Suggestions?

  • Moderator

    @trebor Its been a while, just to be clear the target system is a linux OS?

    If so, pxe boot back in debug mode and see if you can mount either /dev/md126 or /dev/md127 and confirm that your files exist. Also it would also be valuable to capture an output from lsblk to see what FOS sees.

  • Sorry it has been a while. Had time to do as I was told and it still will not image.

    The machine has five (5) drives. Two are the OS on raid1 (mirror) and the other three are a LVM volume of data that I do not have any interest in imaging.

    Read thru the https://forums.fogproject.org/topic/7882/capture-deploy-to-target-computers-using-intel-rapid-storage-onboard-raid

    The raid device FOG is seeing as “/dev/md126” (and /dev/md127) of the OS drives.

    “cat /proc/mdstat” reports ‘good’ raid1.

    Set host to be “captured” and the image fails “Could not find partitons (bin/fog.upload)” and will reboot in one minute.

    I deleted the host and re-inventoried it and set “Host Kernel Arguments” to ‘mdraid=true’ and “Host Primary Disk” to ‘/dev/md126’ but it still fails with same waring.


  • Moderator

    @trebor I guess the simple fact is this:
    Yes we are behind moving documents in the tutorial (instructions and discussions) to the wiki. We could always use more hands to help with the documentation processes. The second part is based on demand. Not a lot of people are asking about software raids in FOG. But you are right, it would be easier if everything was in the wiki. That I can’t dispute.

  • @trebor So, I’m new to FOG and not familiar with all the related terms. Maybe I searched for the wrong terms but, I have to ask, why isn’t this information (https://forums.fogproject.org/topic/7882/capture-deploy-to-target-computers-using-intel-rapid-storage-onboard-raid) incorporated into a easy to find wiki page yet? Me, the newbie, maybe the tags “mdadm” should be included? I don’t know, just suggesting.

  • Thanks for the quick reply george1421.

    The page I have been looking for! https://forums.fogproject.org/topic/7882/capture-deploy-to-target-computers-using-intel-rapid-storage-onboard-raid


  • Moderator

    While I haven’t personally imaged a linux computer using a software raid, the same rules should apply as the fake-raid articles. The FOS (Fog Operating System) [the software that runs on the target computer to capture and deploy images] is based on linux. So you have a good chance on getting FOS to copy your linux image. You do need to set the kernel parameter mdraid=true this tells FOS to load the raid kernel drivers. You should set that in the host definition for that computer. The next thing you will need to set is what is the raid array device known by inside linux? Its going to be know by /dev/md0? If that is its device ID then that is the value you key in to the host definition. Its important to know the device definition is what FOS sees not your target’s host OS.

    One way to debug this is to:

    1. Manually register this host.
    2. Set the kernel parameter mdraid=true
    3. Schedule a capture or deploy (for this part it doesn’t matter which) but before you hit the submit button check the option for debug.
    4. Now pxe boot your target computer.
    5. After a few enter key presses you will be dropped to a command prompt on the target computer.
    6. Key in lsblk and review what FOS sees.

    That should tell you what device FOS is seeing the software raid as. Key that into the host definition.

    Ref: https://forums.fogproject.org/topic/7882/capture-deploy-to-target-computers-using-intel-rapid-storage-onboard-raid