I came across similar threads while searching for an answer to this problem, but no resolutions.

Running SuSE Enterprise Server 10 on a server that has been in production and booting without problems. I haven't changed the SuSE installation which resides on /dev/sda, but I recreated /dev/sdb after putting in larger disks and rebuilding that array. Since I've done that the server "hangs" on "Importing Net File System (NFS)", then finally returns with "mount server reported tcp not available, falling back to udp". There is no firewall running on this server.

When the server finally comes up it has not mounted a NFS share that's in /etc/fstab, but I can mount it without delay or error from the command line. This server has a twin with a nearly identical configuration that doesn't exhibit these problems, or at least it didn't on last boot.

Any suggestions where I should be looking? I've included what information I could find that I thought was relevant, but please let me know if I've overlooked something.

Thanks - Paul

remus:/etc # cat SuSE-release
SUSE Linux Enterprise Server 10 (x86_64)
VERSION = 10


remus:/etc/init.d # more /var/log/boot.msg
<SNIP>

<notice>portmap start
Starting RPC portmap daemon<notice>startproc: execve (/sbin/portmap) [ /sbin/portmap ], [ CONSOLE=/dev/console ROOTFS_FSTYPE=ext3 TERM=linux SHELL=/bin/sh ROOTFS_FSCK=0 LC_ALL=POSIX INIT_VERSION=sysvinit-2.86 REDIRECT=/dev/tty1 COLUMNS=155 PATH=/bin:/usr/bin:/sbin:/usr/sbin vga=0x31a RUNLEVEL=3 PWD=/ SPLASHCFG=/etc/bootsplash/themes/SuSE-SLES/config/bootsplash-1280x1024.cfg PREVLEVEL=N LINES=60 SHLVL=2 HOME=/ splash=silent SPLASH=yes ROOTFS_BLKDEV=/dev/sda3 _=/sbin/startproc DAEMON=/sbin/portmap ]
done
<notice>'portmap start' exits with status 0
<notice>splash_early start
<notice>'splash_early start' exits with status 0
<notice>nfs start
Importing Net File System (NFS)mount server reported tcp not available, falling back to udp
mount: RPC: Remote system error - Connection timed out
done
<notice>'nfs start' exits with status 0
<notice>nfsboot start
Starting nfsboot (sm-notify) done
<notice>'nfsboot start' exits with status 0

<SNIP TO END>

remus:/etc # view /var/log/messages
<SNIP>

Oct 13 08:06:26 remus sshd[3395]: Server listening on :: port 22.
Oct 13 08:06:26 remus zmd: NetworkManagerModule (WARN): Failed to connect to NetworkManager
Oct 13 08:06:26 remus /usr/sbin/cron[3497]: (CRON) STARTUP (V5.0)
Oct 13 08:06:26 remus kernel: Installing knfsd (copyright (C) 1996 okir@monad.swb.de).
Oct 13 08:06:26 remus kernel: NFSD: Using /var/lib/nfs/v4recovery as the NFSv4 state recovery directory
Oct 13 08:06:26 remus kernel: NFSD: starting 90-second grace period
Oct 13 08:06:27 remus xinetd[3574]: Reading included configuration file: /etc/xinetd.d/chargen [file=/etc/xinetd.conf] [line=26]
Oct 13 08:06:27 remus xinetd[3574]: Reading included configuration file: /etc/xinetd.d/chargen-udp [file=/etc/xinetd.d/chargen-udp] [line=14]
Oct 13 08:06:27 remus xinetd[3574]: Reading included configuration file: /etc/xinetd.d/cups-lpd [file=/etc/xinetd.d/cups-lpd] [line=15]
Oct 13 08:06:27 remus xinetd[3574]: Reading included configuration file: /etc/xinetd.d/cvs [file=/etc/xinetd.d/cvs] [line=11]
Oct 13 08:06:27 remus xinetd[3574]: Reading included configuration file: /etc/xinetd.d/daytime [file=/etc/xinetd.d/daytime] [line=16]

<SNIP TO END>


remus:/etc/init.d # rpcinfo -p localhost |grep nfs
100003 2 udp 2049 nfs
100003 3 udp 2049 nfs
100003 4 udp 2049 nfs
100003 2 tcp 2049 nfs
100003 3 tcp 2049 nfs
100003 4 tcp 2049 nfs


remus:/etc/init.d # chkconfig |egrep -i 'nfs|firewall'
SuSEfirewall2_init off
SuSEfirewall2_setup off
nfs on
nfsboot on
nfsserver on

I'm replying to this because it is 6 days old and there has been no reply rather than because I know much about it but I did research a similar problem for what turned out to be tardy or failing network setup by VirtualBox 3.early for a guest NFS-mounting /usr from the host. I know nothing about SuSE's boot script mechanism.

As you have probably guessed, the problem is fundamentally about the network not being set up when the mount attempt starts; there were a couple of problems discussed on the 'net that turned out to be caused by a tardy Cisco device and cured by update.

Alternatively you may have some success by simply leaving the NFS mounts until later. Slackware boot scripts do this by initially excluding nfs (amongst others) from the initial mount -a by adding -t nonfs and then mounting the NFS file systems listed in fstab later in the boot process.

Any such workaround does not identify why your system used to work OK and now does not

Good fortune!

Thank you for your reply... I appreciate the input. I agree that it seems like the network was failing. I realized in re-reading this post that I hadn't included the network start section of the boot.msg. The network appeared to be starting okay but maybe there was some latency in actually being able to pass network traffic prior to the NFS client trying to come up.

As a follow-up, the problem does appear to be resolved but I never identified why it was failing. I was also having trouble getting the registration activation code uploaded to Novell so we could get the latest service packs loaded; from looking around on the Web it looks like Novell/SuSE Enterprise registration has been problematic for a number of people. Last week I got the activation code registered with Novell and brought the server up to SLES 10 SP2. That appears to have fixed the problem.

Again, I appreciate your feedback, and I hadn't considered that the problem could have been with latency at the switch (which is a Cisco). That's something I will check if I encounter similar problems with any network services in the future.

Thanks again - Paul