Fog taking forever to boot to the network


  • Hi, I have a new fog installed and when we boot to the network, it sits at mtftp for like 5 minutes before we get the fog boot menu. I ssh in and it says the Tito server is inactive/dead so I restart it and it works fine for a few minutes. But then we get back to waiting at mtftp for 5 minutes again. What is happening and how do I fix it? It’s running on a server 2016 hyper-v on centos 7

  • Moderator

    @anwoke8204 said in Fog taking forever to boot to the network:

    tftp 10.4.47.15 ipxe.efi c:\users\kstevens

    Hopefully you just missed the action command when you posted the command.

    The command should be
    tftp 10.4.47.15 get ipxe.efi c:\users\kstevens and it should place the file ipxe.efi in the directory path you provided. It concerns me that the destination file name did not match the source name. The transfer should be very quick like about 2-3 seconds from the time you hit enter on that get command.

    What are ping times to this fog server 10.4.47.15 from this same windows test computer?

    What are ping time if you use this command ping -4 -l 1300 -n 4 10.4.47.15 This command says use ipv4 set the payload size to 1300 and ping 4 times to the host. I’m specifically using a payload size larger that the default ping size to see if we are hitting a mtu issue that wouldn’t be visible with a 64 byte ping size.


  • hi, when I do that, it gives me a file called TFTP1168. it takes a bit to connect, but then gives me that file. I am using the command tftp 10.4.47.15 ipxe.efi c:\users\kstevens

  • Moderator

    @anwoke8204 Ok so we can ignore a half a do this test.

    The issue seems to be with tftp or dhcp if you look at the truth table.

    So if you take a windows computer on the same IP subnet as the failing pxe booting computer, install the tftp client feature. Drop the windows firewall. Then use the windows tftp client to get ipxe.efi from the FOG server’s IP address. I’m interested in seeing it transfer the file, and the transfer happens quickly. The ipxe.efi file is pretty small so it should transfer fast.


  • @george1421 correct

  • Moderator

    @anwoke8204 So just for clarity when it take 5 minutes to pxe boot and ssh says the tito server is dead you can still access the fog web ui without issue?


  • @george1421 the server is active and running just fine. We can get into the web interface, ping and everything just fine, and I check the network and don’t find any issues with the network either.

  • Moderator

    @anwoke8204 Is the centos vm failing under hyper-v? When ssh says the server is inactive/dead what do you see on the hyper-v console. Is the server still functional?

    If the server itself is still functional but it drops off the network, I’ve seen something similar to your explanation when there are two devices on your subnet with the same IP address. Each fight over ownership of the IP address. This makes communication work right after a reboot but then fail sometime later when another device advertises the same IP address. I’m not saying that this is the problem for you, just based on what you have mentioned so far, it kind of fits.

289
Online

9.9k
Users

16.2k
Topics

149.0k
Posts