Results 1 to 5 of 5
  1. #1
    Join Date
    Oct 2003
    Location
    Hanoi
    Posts
    4,306

    tftp - Transfer timed out

    I am setting up tftpd server for PXE usage, but when testing from tftp client, it always shows Transfer timed out, could not get any file. /var/log/messages shows client is connected already.

    UDP 69 is open.
    Any idea please?

    Thank you.

  2. #2
    Join Date
    Mar 2009
    Location
    Chicago, IL
    Posts
    219
    Tcpdump is your friend here. Also if you could tell us a bit more about your setup that would be very helpful.

    If these are standard servers plugged into normal switches, one thing to keep an eye out for is spanning tree implementations on the switches. Generally when a machine first powers on and the port detects link, it'll block for a bit until it can "figure out" what's going on and then bring the port fully up in a non-blocking mode. If the time spent in blocking mode is greater than your timeout on tftpd, then that could be a big contributor.

    Also, you should ensure that the NICs are set to be your primary boot devices in the BIOS of the machines you're using. That way you can skip the time spent waiting for the HDs/CDRoms to poll looking for a valid boot device.

    If these switches are cisco based, you can do something like this for their particular interface to disable spanning tree...

    Code:
    conf t
    int fa0/XX (xx is the port number, also assuming fast ethernet here)
    no spant
    ^z
    wr mem
    That should get you setup with an immediate link-up on your port. Just *DO NOT CROSSLINK THIS PORT INTO ANOTHER SWITCH*. Routing loops are bad, mmkay? =P

  3. #3
    Join Date
    Oct 2003
    Location
    Hanoi
    Posts
    4,306
    No, spanning tree is not activated. Server was rebooted from NIC. It can take dynamic IP fine. Here is the config of tftpd

    service tftp
    {
    socket_type = dgram
    protocol = udp
    wait = yes
    user = root
    server = /usr/sbin/in.tftpd
    server_args = /tftpboot
    disable = no
    per_source = 11
    cps = 200 5
    flags = TCPv4
    }

  4. #4
    Join Date
    Oct 2003
    Location
    Hanoi
    Posts
    4,306
    to follow this, here is tcpdump result

    # tcpdump -vvv port 69
    tcpdump: listening on eth0, link-type EN10MB (Ethernet), capture size 96 bytes
    17:05:55.565442 IP (tos 0x0, ttl 50, id 53077, offset 0, flags [none], proto: UDP (17), length: 49) client-ip.23021 > server-ip-rdns.tftp: [udp sum ok] 21 RRQ "tests.txt" netascii
    17:06:00.564860 IP (tos 0x0, ttl 50, id 62293, offset 0, flags [none], proto: UDP (17), length: 49) client-ip.23021 > server-ip-rdns.tftp: [udp sum ok] 21 RRQ "tests.txt" netascii
    Is there any idea?
    Thanks.

  5. #5
    The issue looks like due to packet loss.

    The better solution is to check the packets flow vis TCPdump and additionally run ifconfig and see if any packets are dropped via NIC

Similar Threads

  1. TFTP Recommendations (windows)
    By HostExchange in forum Hosting Security and Technology
    Replies: 2
    Last Post: 11-22-2007, 06:06 PM
  2. Cisco PIX 515 - Best way to restore image via TFTP
    By Georgecooldude in forum Hosting Security and Technology
    Replies: 1
    Last Post: 04-24-2006, 01:27 PM
  3. Tftp/fedora3/transfer Timed Out
    By coolboarderguy in forum Hosting Security and Technology
    Replies: 2
    Last Post: 02-22-2005, 11:17 PM
  4. tftp & tftpd ??
    By brands in forum Hosting Security and Technology
    Replies: 3
    Last Post: 12-19-2002, 04:16 AM
  5. timed out
    By clicker in forum Web Hosting
    Replies: 3
    Last Post: 12-11-2000, 02:55 AM

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •