Subject: misc/12279: netboot client hangs in /etc/rc.d/root
To: None <firstname.lastname@example.org>
From: Ingolf Steinbach <email@example.com>
Date: 02/24/2001 23:11:12
>Synopsis: netboot client hangs during boot in /etc/rc.d/root
>Arrival-Date: Sat Feb 24 14:13:00 PST 2001
>Originator: Ingolf Steinbach
System: NetBSD isdn 1.5 NetBSD 1.5 (ISDN) #0: Fri Feb 23 21:41:15 MET 2001 ingolf@isdn:/usr/obj/sys/arch/i386/compile/ISDN i386
When netbooting, the boot process hangs during execution of
/etc/rc.d/root. These are the last words:
root on we1
nfs_boot: trying RARP (and RPC/bootparam)
nfs_boot: client_addr=192.168.2.4 (RARP from 192.168.2.1)
root on boot:/export/isdn/root
Automatic boot in progress: starting file system checks.
mount_nfs: rpcbind on server: RPC: Unable to send
This happens in the "mount /" line. The "Unable to send" message
is repeated after a while.
The machine has already successfully got its network configuration
via rarp and bootparam. The remote root directory is already
mounted via nfs at this point (i.e. it is not unmounted by the
preceeding "umount -a >/dev/null 2>&1").
In case this is important: both (netboot-) server and client
run kernels with IPv6 enabled. The client's site local v6-address
is not yet configured at this point.
A tcpdump on the server does not show any traffic during the
"Unable to send" messages.
After hitting ^C on the client, the boot process continues
Netboot a machine (with IPv6 kernel). Watch it hang during
At the moment, I am using the following workaround, but this
probably is not The Right Thing:
# echo 'rcvar=root' >> /etc/rc.conf.d/root
# echo 'root=NO' >> /etc/rc.conf