Diskless boot problem

[Sławek Żak]

On 5/22/05, Doug White <dwhite at gumbysoft.com> wrote:
> On Sat, 21 May 2005, [ISO-8859-2] S�awek �ak wrote:
> 
> > On 5/20/05, Doug White <dwhite at gumbysoft.com> wrote:
> > Please try to avoid sending your message bodies base64 encoded :) My mail
> > client got really confused by it and didn't quote the message properly.
> >
> > Also stripping hackers cc:.
> 
> I'd like to, but Gmail doesn't offer this option. Sorry :(
> 
> > On Thu, 19 May 2005, [ISO-8859-2] S³awek ¯ak wrote:
> >
> > > Hi,
> >
> > > > I have a problem with booting Dell 2850 over network. The machine reads
> > > > kernel over net, boots upto mounting / from NFS and then crashes.
> > >
> > > What is the NFS server? It seems to think the NFS handle we pulled the
> > > kernel with is no longer valid.
> 
> > FreeBSD 5.3/5.4-STABLE.
> 
> Hm ... dunno then ... does the server complain about the client at all in
> the log?

No complaints whatsoever.The exact os version on NFS server is 5.4-RC2.

> 
> >> Does PXE and the system itself end up pulling different IP addresses?
> 
> > No. The IP is the same.

Message on the console before boot:

CLIENT MAC ADDR: 00 11 43 D3 6E 09  GUID: 44454C4C 4600 1038 8031 B9C04F44314A
CLIENT IP: 10.158.190.74  MASK: 255.255.255.224  DHCP IP: 10.158.190.73
GATEWAY IP: 10.158.190.94                                                     
PXE Loader 1.00                                                        
                         
Building the boot loader arguments
Relocating the loader and the BTX
Starting the BTX loader

BTX loader 1.00  BTX version is 1.01

pjd@ sent me a patch for hardcoding 100 full-duplex on em, but it didn't help:

em_init_locked: Forcing 100Mbit/full-duplex
NFS ROOT: 10.158.190.73:/var/www/FreeBSD-5.4-x86-PXE
em0: Link is up 100 Mbps Full Duplex
exec /sbin/init: error 70
exec /sbin/oinit: error 70
exec /sbin/init.bak: error 70
exec /rescue/init: error 70
exec /stand/sysinstall: error 70
init: not found in path
/sbin/init:/sbin/oinit:/sbin/init.bak:/rescue/init:/stand/syl
panic: no init
cpuid = 0
KDB: enter: panic
[thread pid 1 tid 100002 ]
Stopped at      kdb_enter+0x30: leave   
db> 

tcpdump trace on the server still shows stale handle NFS requests.

12:16:09.343817 arp who-has 10.158.190.73 tell 10.158.190.74
12:16:09.343828 arp reply 10.158.190.73 is-at 00:11:43:d3:6e:e1
12:16:09.345688 IP 10.158.190.74.1830074133 > 10.158.190.73.2049: 100
lookup [|nfs]
12:16:09.345721 IP 10.158.190.73.2049 > 10.158.190.74.1830074133:
reply ok 28 lookup ERROR: Stale NFS file handle
12:16:09.351560 IP 10.158.190.74.1830074134 > 10.158.190.73.2049: 100
lookup [|nfs]
12:16:09.351579 IP 10.158.190.73.2049 > 10.158.190.74.1830074134:
reply ok 28 lookup ERROR: Stale NFS file handle
12:16:09.386539 IP 10.158.190.74.1830074135 > 10.158.190.73.2049: 100
lookup [|nfs]
12:16:09.386560 IP 10.158.190.73.2049 > 10.158.190.74.1830074135:
reply ok 28 lookup ERROR: Stale NFS file handle
12:16:09.422644 IP 10.158.190.74.1830074136 > 10.158.190.73.2049: 100
lookup [|nfs]
12:16:09.422663 IP 10.158.190.73.2049 > 10.158.190.74.1830074136:
reply ok 28 lookup ERROR: Stale NFS file handle
12:16:09.460996 IP 10.158.190.74.1830074137 > 10.158.190.73.2049: 104
lookup [|nfs]
12:16:09.461019 IP 10.158.190.73.2049 > 10.158.190.74.1830074137:
reply ok 28 lookup ERROR: Stale NFS file handle
12:16:09.498724 IP 10.158.190.74.1830074138 > 10.158.190.73.2049: 104
lookup [|nfs]
12:16:09.498745 IP 10.158.190.73.2049 > 10.158.190.74.1830074138:
reply ok 28 lookup ERROR: Stale NFS file handle


I'm out of options.

Thanks, /S
-- 
Sławek Żak / UNIX Systems Administrator