Strange NFS root booting behavior.

Vladimir Botka vlado at
Thu Apr 5 06:34:02 UTC 2007

> I've set up a dhcp server to support two (or more) clients, which
> mount their root on
> an nfs shared volume.
> For one client, the booting process has one symptom:
> after the initial PXEBOOT and getting the parameters from the dhcp 
> server, the
> boot process hangs at:
> pxe_boot: gateway ip: ...
> And after a few minutes eventuall loads the loader.conf
> and after a few more minutes, loads the kernel, which seems pretty
> speedy given the twirlers.
> On the other system, oddly, the boot process seems to whip through the
> above boot section with no hesitation. However, after the kernel is
> loaded, and the init starts up, with in a ver few initial commands
> there is a diagnostic to the effect:
> Starting file system checks:
> Then ...
> mount /: Bad file descriptor
> Mounting root filesystem rw failed, startup aborted.
> Since I'm booting off of the same hierarchy, I did create a ramdisk
> and initialized
> a file system to have '/var/run' created local to the client systems.
> Are there other things that need to be separated for each client? And
> would failing to do set these up, give the above behavior, in
> particular the failure to boot
> and run through the normal rc scripts.
> Thanks
> John Clark.
> _______________________________________________
> freebsd-current at mailing list
> To unsubscribe, send any mail to
> "freebsd-current-unsubscribe at"


try the script /usr/share/examples/diskless/clone_root to create copy
of the root. The procedure in
works for me. Just an idea but did you compile the diskless kernel with
"options     BOOTP_NFSROOT" ?

Cheers, -vlado

More information about the freebsd-current mailing list