V0.32 Storage Node Selection Process?

  • I am on V0.32. I tried 1.2, but it had major issues, most likely do to the upgrade process, and had to rollback.

    My question is that we just have started using nodes at our remote site, so we can image on site, and not pull the image over the small MPLS connection. I have setup the node and everything seems to be working fine.

    However, I have run across instances where computers I am imaging are pulling from the remote site.
    Reading through the wiki and forums, there is no logic I can find that specifies how a specific imaging job selects which storage node it pulls the data from.

    As far as I can tell, its random. If I image 2 computers at the same time, the first images off the HQ Node, the second images off the remote Node.

    Shouldn’t it be looking for the closest node, and select that? Is there a way I can specify the node per task or by editing the PXE boot menu?

    For now I am disabling the remote node until we need to image a computer at the remote site. Then I enable the remote node and disable all others. This seems to be a bit of administrative overkill…

    Network topography is as follows:
    FogServer (main node), vlan 1 (server vlan)
    –Gigabit Switches–
    HQ Node (Secondary Node) Vlan 10 (PC Vlan)
    –Gigabit Switches –
    Worksation Imaging Area VLAN 24
    –3x3MPLS Network–
    Remote Site Note - VLAN 45 (remote site PC VLan)
    Remote Site Workstation - VLAN 45

    Any advice is appreciated!

  • The OP’s problem can be solved with the location plugin in FOG 1.2.0, and FOG 1.3.0. Documentation on the plugin are linked below. Marking this thread as solved.

  • @HABCO we have tutorials for installing from scratch. It doesn’t take more than maybe 3 hours. Most of that time is waiting for updates to run.

  • I originally deployed FOG using the VM template for version 0.27 and upgraded from there when I started using it 4 years ago.

    Has anyone developed a FOG 1.20 VM Template? I’m fine managing linux servers for specific use, but building from scratch is something I don’t have time for right now.

    Also, a new install erases all our history, and we use fog for a pseudo inventory management as well. I’ll try booting my TRUNK version I had upgraded to a few weeks ago, and do a re-image upload. The big problem I had was that the images wouldn’t deploy. maybe if I do a fresh image it would solve that.


  • Moderator

    @HABCO I would recommend that if you want to upgrade to the trunk version to install it clean on a new server and not upgrade from 0.30. This will give you two positives.

    1. You will not touch your current fog installation
    2. Upgrading 0.30 sometimes has difficulties with the upgrade that end up taking longer to fix then just deploying clean.

    You can copy your current images from the 0.30 fog server to the 1.2.0 trunk server and then just manually key in the image information to complete the migration. Then just update your dhcp option 66 and 67 settings for your new install. If something goes horribly wrong then you can always change the dhcp settings back and resume your existing fog 0.30 environment.

    In the end Wayne is absolutely right you need to be on version 1.2.0 or newer to do what you want with the location plugin.

  • I’d be more than happy to help you migrate to 1.2.0, or to FOG Trunk.

    Not certain on this but the location plugin probably wasn’t thought of in 0.32, which is exactly what you need. It’s available in 1.2.0 and in FOG Trunk. We have a wiki article about it, but it covers it from the perspective of the soon-to-come FOG 1.3.0, which is currently being developed still (June 2016), and known as FOG Trunk. https://wiki.fogproject.org/wiki/index.php?title=Location_Plugin It’s possible to upgrade to FOG Trunk from any version of FOG, but I advise to deploy all your images to physical machines (but not boot them), upgrade, and then re-upload all the images to Trunk because there are differences in how they are handled and stored in trunk.

    Increasingly, many people use FOG Trunk in production because of it’s features, UEFI support, GPT support, Win10 Support, Security enhancements, and massive performance improvements. Find links to it here: https://wiki.fogproject.org/wiki/index.php?title=Upgrade_to_trunk You will need Ubuntu 14 or higher, or Fedora 21 or higher, or CentOS 7, or Debian 8. Don’t bother if you are on a super-old OS, just build a new server using a new OS and then simply migrate images one at a time by downloading/uploading.