UNSOLVED Failed to get an ip from DHCP ! Tried on interfaces eth0 eth0 .

  • Server
    • FOG Version: 1.3.5-RC-16
    • OS: Debian 8
    • Service Version:
    • OS:

    So my problem is that suddenly fog is not able to bring up the interfaces on machines and get ips . It boots up the pxe i select Full inventory and Registration then it loads some files but always fails at : Starting eth0 inerface blablabla . Also before that i am gatting an error on Populating /dev using udev : error creating epoll fd : Function not implemeted. This is very strange everything worked fine until suddenly today i am having this issue. Check on differents switches but nothing . also different devices same error. Can anyone help ?

  • While I doubt this will fix the problem, I do hope to have a nicer approach to having a device receive an IP Address now.

    As stated before, the current mechanism just waited for “ANY” ip address to be assigned to any NIC as needed. As long as it received an IP address (regardless of if that IP could communicate to the FOG Server) it would break out and try “driving” forward with the tasking.

    I’ve added what I hope will allow a more dynamic approach to detecting the real fog network. Would you mind giving a try using the init’s and instructions here:

    wget -O /var/www/fog/service/ipxe/init.xz http://mastacontrola.com/init.xz
    wget -O /var/www/fog/service/ipxe/init_32.xz http://mastacontrola.com/init_32.xz

    With this, you’ll also get the latest fixes from last night in regards to resizing code bits and what not. (This isn’t a bad thing)

  • Moderator

    @kleanthis said in Failed to get an ip from DHCP ! Tried on interfaces eth0 eth0 .:

    New Switches were installed that only have FastEthernet

    ?? FastEthernet == 100Mhz ??

  • Moderator

    @kleanthis said in Failed to get an ip from DHCP ! Tried on interfaces eth0 eth0 .:

    @Tom-Elliott Well why cant my lan 1 and 2 get an ip but lan 3 can . This is my actual question . Otherwise when an os is installed all lan ports work fine .

    (sorry working at my real job right now).

    What I find stange is that the link light is not coming on, on this lan1 lan2 interface. The link light is a hardware/software function. If there is a functioning network jack, the light should come on.

    From the FOS side, its seeing the network adapters AND talking to them since its getting a mac address from the network adapters. So FOS is happy with the network adapters.

    I might understand why they are not getting an IP address depending on how FOS manages the network adapters. We may need to issue the dhcp command again on these interfaces (once we can understand why they link light is not coming on).

  • @kleanthis The “old environment” is completely separate from the now new “FastEthernet” system?

  • New Switches were installed that only have FastEthernet , but because i thought of that as well i tested it in my old environment Gigabit downstairs were it was working before. and still same issue.

  • What’s even more boggling, to me, is your post that it just “started” happening. “Everything worked fine until suddenly today I am having this issue.”

    Did something in your networking environment change?

  • @kleanthis I don’t know, that’s what we’re trying to get to the bottom of.

    There are any number of reasons why, and it seems (to me) that possibly Spanning tree, or Power Saving features are preventing things from “normally” happening.

  • @Tom-Elliott Well why cant my lan 1 and 2 get an ip but lan 3 can . This is my actual question . Otherwise when an os is installed all lan ports work fine .

  • @kleanthis this is kind of by design, as @george1421 was saying earlier.

    It’s not caring “what” the IP address is. The idea is, get an IP address, and as soon as one is available continue forward with the tasking. Otherwise you’re waiting for each interface to come up before your tasking can start. (This can be a bad thing if there is nothing for the device to pick up).

  • @george1421 Nope , the other two lan ports dont get an ip . Their leds dont even come on . Its like they are unplugged or not activated. As soon as i plug the cable on lan 3 lights come up.

  • Moderator

    @kleanthis OK now we know the order that FOS is seeing the interfaces.

    eth0: 00:0b:ab:70:9a:21 (
    eth1: 00:0b:ab:70:aa:69 
    eth2: 00:0b:ab:70:aa:6a

    If this was a typical “server” I would say that eth0 is considered an add on network adapter (which is typically discovered first) and then the built in network adapters (eth1 and eth2). Where eth1 and eth2 mac address are consistent with a dual port network adapter. What’s happening is understandable.

    For the sake of discussion, if you plug in the other two nics and run the ip addr show again, do the other interfaces get an IP address too? You may loose access via putty when you do that, but you can still key things in on the FOS console.

    The bigger question now is why isn’t FOS testing each network interface for an up link. Right now it sounds like its giving up after testing eth0 (LAN3) and not continuing to loop.

  • There you go :

    [Fri Mar 17 root@fogclient ~]# ifconfig
    eth0      Link encap:Ethernet  HWaddr 00:0B:AB:70:9A:21
              inet addr:  Bcast:  Mask:
              UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
              RX packets:1600 errors:0 dropped:0 overruns:0 frame:0
              TX packets:67 errors:0 dropped:0 overruns:0 carrier:0
              collisions:0 txqueuelen:1000
              RX bytes:183345 (179.0 KiB)  TX bytes:15605 (15.2 KiB)
    lo        Link encap:Local Loopback
              inet addr:  Mask:
              UP LOOPBACK RUNNING  MTU:65536  Metric:1
              RX packets:0 errors:0 dropped:0 overruns:0 frame:0
              TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
              collisions:0 txqueuelen:1000
              RX bytes:0 (0.0 B)  TX bytes:0 (0.0 B)
    [Fri Mar 17 root@fogclient ~]# ip addr
    1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue qlen 1000
        link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
        inet scope host lo
           valid_lft forever preferred_lft forever
    2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast qlen 1000
        link/ether 00:0b:ab:70:9a:21 brd ff:ff:ff:ff:ff:ff
        inet brd scope global eth0
           valid_lft forever preferred_lft forever
    3: eth1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop qlen 1000
        link/ether 00:0b:ab:70:aa:69 brd ff:ff:ff:ff:ff:ff
    4: eth2: <BROADCAST,MULTICAST> mtu 1500 qdisc noop qlen 1000
        link/ether 00:0b:ab:70:aa:6a brd ff:ff:ff:ff:ff:ff

  • @kleanthis This connecting via putty is nice because you can copy/paste the contents of the output from commands we might request.

  • Moderator

    @george1421 When you get to the FOS command line, if you give root a password, you can connect to the FOS engine via putty to help with any copy paste operations you need.

    at the FOS command line key in ip addr show to get the device’s IP address
    then key in passwd and give root a password. It doesn’t matter what password you give it, as long as its not blank. From there you can connect to the FOS engine using putty and login as root and what ever password you give it. Then you can run the debugging commands from there.

  • Moderator

    @kleanthis what we “need” is to get to the linux command prompt on the target computer THEN key in ip addr show, or ip link show. From the output then we can see the reference between ethX and mac address. That will help us identify what FOS “thinks” eth0 is.

    Don’t worry about breaking what you were just able to do, we are NOT going to start the deploy/capture as long as debug was enabled.

  • @george1421 ok so use lan 3 (the one thats working) for the debug deployment?

  • Moderator

    @kleanthis Its simple. Just schedule another capture/deploy to the same target. Since you have LAN3 working use that interface. But before you submit the task select the check box that says debug.


    This tells FOS to not start the task automatically but drop the user to a linux command line. This is what the developers use to debug the FOS engine.

  • Is there a guide on this ? Debug deployment?

  • Moderator

    @kleanthis it would be interesting (again) to setup a debug deploy/capture to get access to the FOS engine command line. I’m suspecting that FOS is setting the eth0, eth1, eth2 order based on hardware discovery order, which seems to be inconsistent with what the iPXE kernel is doing.