Find the answer to your Linux question:
Results 1 to 5 of 5
Hello, I have recently set up 2 NFS Servers (Server A and Server B) each hosting home directories. Users are able to mount their NFS home directory shares via ypbind ...
  1. #1
    Just Joined!
    Join Date
    Mar 2008
    Posts
    24

    NFS Trouble

    Hello,

    I have recently set up 2 NFS Servers (Server A and Server B) each hosting home directories. Users are able to mount their NFS home directory shares via ypbind and autofs. When I am on my computer I can log into any other machine on the network. Essentially my home directory mounts fine on the rest of the networked systems.

    HOWEVER, when I try to log into Server B (my home directory is on Server A) I cannot. I get the following error on Server B in /var/log/messages:

    kernel: nfs: server A not responding, timed out

    THEN on Server A, in /var/log/messages, I get:
    mountd: authenticated mount request from IP of Server B.

    I don't get it. Server A is receiving Server B's request to mount my home directory and says it's authenticating my then Server B responds that it Server A is not responding.

    I have verified on both systems:
    nfs on
    portmaps on
    iptables off

    How is it that I can log into any other machine on the network except the new Server B??

    I'm stuck! Please help!

    Starr

  2. #2
    Just Joined!
    Join Date
    Jan 2007
    Location
    San Antonio, TX
    Posts
    21
    Eliminating some of the more advanced/automated options can you do a simple NFS mount on server B from server A?

  3. #3
    Just Joined!
    Join Date
    Mar 2008
    Posts
    24
    Not able to hard mount at all either on server B from server A and vice versa doing:

    mount serverB:/phd/home/user /home/user
    mount serverA:/phd/home/user /home/user

    It just times out.

  4. #4
    Just Joined!
    Join Date
    Mar 2008
    Posts
    24
    Just an update: I have resolved the problem. It wasn't anything I expected it to be: The MTU size on the home directory servers were set for 9000 while the rest of the network AND the switches that connected the home directory servers together is set to process MTUs of 1500.

    The problem was that the switches were dropping the packets as soon as the mount requests between the home directory servers were made since the switches were at a default of 1500 MTU and not set to process jumbo packets of 9000 MTU.

    When I did a tracepath from one home directory to another it would get hung up versus when I did a tracepath from a user directory server (1500 MTU) to the home directory server (9000 MTU) which did not. It was then I noticed the different MTU sizes.

    A simple ifconfig change of the MTU size from 9000 to 1500 on the home directory servers immediately fixed the nfs mounts between them.

  5. #5
    Just Joined!
    Join Date
    Mar 2009
    Posts
    6
    Did you run the xinetd and dhcpdon your host machine and dhcpd shoud containg your a and b machines ip addresses.

Posting Permissions

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