Configure fog server to work with subnets



  • Hello all,

    I got a fog server working on ubuntu with 2 nic’s and 2 ip’s the 10.10.200.13 and the 10.11.200.6. On both subnets, pxe boot’s ok but i can only register machines and deploy tasks if the fog server is configured to a specific subnet. For example If i try to register a machine with the ip 10.11.200.* and if the fog_wol_host, tfthost etc is configured to 10.10.200. it will freeze after attempting to send hardware info. Summing it up how can i configure fog to accept requests of both subnets.

    P.S.- oh and the fog server its not a dhcp server.



  • Hi.

    I’m working with the OP at the same place so I can give you more information:

    We have two separated networks which should be served by a single fog server.

    I have 3 nics on the fog server

    1. Subnet A (10.10.200.)
    2. Subnet B (10.11.200.)
    3. iSCSI Network

    I have 2 DHCP Servers. Both have option 066 and 067 configured. DHCP Subnet A has FOG_IP_SUBNET_A and DHCP Subnet B has FOG_IP_SUBNET_B Fog works perfectly on Subnet A.

    However, today we needed to deploy 84 Machines on subnet B and even though I had access to the webmanagement (through NIC B) and also to the PXEBOOT (as assigned through DHCP B), I couldn’t do neither registrations nor deploy the images. To solve the issue (working against Time), I changed all the configs in the management window and FOG started to operate exclusively on Subnet B. I managed to deploy all 84 Images without problems. Now I would like not to have to change manually everything each time I need to deploy an image in diferent Subnets.

    On the old forum I found a topic dealing with this were the OP changed iptables’s configuration to forward requests to NIC A but I’m not sure if that was all he did to accomplish what we also want to do. I also have serious doubts about the new DHCP config that has to be made (should both continue to point to the correct interfaces or should I point DHCP B it to Fog’s Subnet A NIC)?

    here what he did:

    [code]
    /usr/sbin/iptables -A FORWARD -i eth0 -d 192.168.1.15 -j ACCEPT
    /usr/sbin/iptables -A FORWARD -s 192.168.1.15 -o eth0 -j ACCEPT
    [/code]

    I also have some doubts about the risks of forwarding packets like that in the Fog Server because our two subnets work like this:

    ---- |-----Internet Gateway A ----- Subnet A ----------|FOG SERVER NIC A
    |
    | > (limited comunication possible here.
    | Out of our reach and not suitable for
    | Image transfer/deployment
    |
    ---- |-----Internet Gateway B ----- Subnet B ---------- | FOG SERVER NIC B

    So, basically, they arn’t allowed to comunicate inside our scope but Subnet A can comunicate to Subnet B on some protocols using the gateways (which is by design and won’t be changed as it is out of our reach).

    Answering your question, our Switches are all HP Procurves 2650 (Layer 2) and a central HP6100 (which then connects to the gateways) interconnected with gigabit fibre but this seems irrelevant to the case since we need to treat the Subnets as isolated (being able to PXboot from Subnet B using NIC B and accessing the management on both NICs is rather peculiar and I can’t explain it unless both TFTPD and Apache are listening on both interfaces). The bottom problem remains… I can’t register hosts or deploy the image through Subnet B) when FOG is configured as primary on Subnet A. I can however PXBOOT and access the Web management interface

    Multicasting is set on the switches or I wouldn’t be able to deploy the images on both Subnets (just not at the same time without changing Fogs configuration)

    Thank you for any input

    Best Regards

    Tony



  • What kind of switches do you have deployed, and are they configured for a layer 2 or layer 3 network?


Log in to reply
 

456
Online

39.4k
Users

11.1k
Topics

105.4k
Posts

Looks like your connection to FOG Project was lost, please wait while we try to reconnect.