LinHES Forums
http://forums.linhes.org/

Attempting a Neoware Eon 100 as FE
http://forums.linhes.org/viewtopic.php?f=2&t=16321		 Page 1 of 1
Author:  garetjax67 [ Tue Sep 04, 2007 4:55 pm ]
Post subject:  Attempting a Neoware Eon 100 as FE
Hey all, I have a Neoware Eon 100 that I am attempting to get working as a diskless frontend. I have hooked an external usb cd-rom to it, and it works great when you boot from the knoppmyth R5F1 cd. So to make a completely silent front end I am attempting to setup diskless boot. I have ran the script:

knoppmyth_diskless_frontend.bash

But during the second phase of the script I get this message:

Unable to automatically determine the MAC address of the current default frontend by interrogating /var/log/daemon.log. This may be because you have not yet booted the diskless FE. You can either boot the diskless FE and try running /user/local/bin/knoppmyth_diskless_frontend.bash again, or delete /nfsfoot/eon and re-create it later.


If you look at /var/log/daemon.log:
Code:
Sep  4 17:57:59 mythtv dhcpd: DHCPDISCOVER from 00:e0:c5:59:61:c3 via eth0
Sep  4 17:57:59 mythtv dhcpd: DHCPOFFER on 192.168.11.26 to 00:e0:c5:59:61:c3 via eth0
Sep  4 17:58:30 mythtv dhcpd: DHCPDISCOVER from 00:e0:c5:59:61:c3 via eth0
Sep  4 17:58:30 mythtv dhcpd: DHCPOFFER on 192.168.11.26 to 00:e0:c5:59:61:c3 via eth0
Sep  4 17:58:33 mythtv dhcpd: DHCPDISCOVER from 00:e0:c5:59:61:c3 via eth0
Sep  4 17:58:33 mythtv dhcpd: DHCPOFFER on 192.168.11.26 to 00:e0:c5:59:61:c3 via eth0
Sep  4 17:58:38 mythtv dhcpd: DHCPDISCOVER from 00:e0:c5:59:61:c3 via eth0
Sep  4 17:58:38 mythtv dhcpd: DHCPOFFER on 192.168.11.26 to 00:e0:c5:59:61:c3 via eth0


I have verified that the MAC is correct. The DHCP successfully works with other systems. The Eon just sits there with a "Sending DHCP requests . . . "

I am really new at this so anywhere else I can look for logs or other hints? Does anybody know what may be the problem? Anything else I can attempt?
Author:  Kirk [ Tue Sep 04, 2007 6:30 pm ]
Post subject: 
Hrm, looks like the client is being offered the 192.168.11.26 address by the server, but it doesn't like it, it should be sending a DHCPREQUEST immediately. You could try setting up a static IP via dhcp using the the host parameter with dhcpd.conf just to test it out;

Code:
host <hostname> {
                hardware ethernet <ethernet address>;
                fixed-address <hostname>;
}


looking something like;

Code:
host frontend1 {
                hardware ethernet 00:e0:c5:59:61:c3;
                fixed-address 192.168.11.20;
}


Just make sure the fixed-address is outside your range of dynamically allocated address.

Hrm, thinking of that have you allocated enough range for all clients on your network? Probably have otherwise the server wouldn't be offering.

Bad cable perhaps?

Some ideas anyways.

Regards,
Kirk.
Author:  garetjax67 [ Tue Sep 04, 2007 7:50 pm ]
Post subject: 
Thanks for the suggestions Kirk.

I modified the dhcpd.conf file statically setting it to .18. It sort of helped same problem on the client, but more info in the logs:

Code:
Sep  4 20:48:07 mythtv dhcpd: DHCPDISCOVER from 00:e0:c5:59:61:c3 via eth0
Sep  4 20:48:07 mythtv dhcpd: DHCPOFFER on 192.168.11.18 to 00:e0:c5:59:61:c3 vi
a eth0
Sep  4 20:48:09 mythtv dhcpd: DHCPREQUEST for 192.168.11.18 (192.168.11.11) from
 00:e0:c5:59:61:c3 via eth0
Sep  4 20:48:09 mythtv dhcpd: DHCPACK on 192.168.11.18 to 00:e0:c5:59:61:c3 via
eth0
Sep  4 20:48:09 mythtv in.tftpd[5511]: tftp: client does not accept options
Sep  4 20:48:18 mythtv dhcpd: DHCPDISCOVER from 00:e0:c5:59:61:c3 via eth0
Sep  4 20:48:18 mythtv dhcpd: DHCPOFFER on 192.168.11.18 to 00:e0:c5:59:61:c3 vi
a eth0
Sep  4 20:48:20 mythtv dhcpd: DHCPDISCOVER from 00:e0:c5:59:61:c3 via eth0
Sep  4 20:48:20 mythtv dhcpd: DHCPOFFER on 192.168.11.18 to 00:e0:c5:59:61:c3 vi
a eth0
Sep  4 20:48:25 mythtv dhcpd: DHCPDISCOVER from 00:e0:c5:59:61:c3 via eth0
Sep  4 20:48:25 mythtv dhcpd: DHCPOFFER on 192.168.11.18 to 00:e0:c5:59:61:c3 vi
a eth0


The DHCPDISCOVER and DHCPOFFER just continue:

Thanks everyone.
Author:  Kirk [ Tue Sep 04, 2007 9:49 pm ]
Post subject: 
I'm not sure why the client doesn't want to use DHCP, but anyways. The client seems to be setting the IP correctly when connecting to the tftp server so we'll run with that.

Where is the client actually booting up to? It seems you had a good 9 seconds between the last ACK and another DISCOVER so I'm presuming it's actually trying to boot the image?

I had a look through the knoppmyth_diskless_frontend.bash script and I noticed that when it creates the /tftpboot/pxelinux.cfg/default file that it tries to set the client to use DHCP again when loading the kernel.

Code:
APPEND root=/dev/nfs ip=dhcp nfsroot= ...


Try changing that to

Code:
APPEND root=/dev/nfs nodhcp nfsroot= ...


at least that's what we use here at work when booting knoppix over PXE and NFS. Otherwise you could try ip=192.168.11.18, but then you'd be limited to one frontend, or you'd have to create different profiles for each MAC.

Good luck,
Kirk.
Author:  garetjax67 [ Wed Sep 05, 2007 7:56 pm ]
Post subject: 
No luck, but I do have new info. on the Client I get the following: (forgive me having to type this over, can't figure out how to copy/paste from the client.)

Code:
Looking up port of of RPC 100005/1 on 192.168.11.11
portmap: RPC call returned error 101
Root-NFS: Unable to get mountd port number from server, using default
mount: RPC call returned error 101
Root-NFS: Server returned error -101 while mounting /nfsroot/eon
VFS: Unable to mount root fs via NFS, trying floppy.
VFS: Insert foot floppy and press ENTER


Thanks!
Author:  Kirk [ Wed Sep 05, 2007 8:43 pm ]
Post subject: 
Error 101 means the network is unreachable.

I'm thinking something's up with your server.. is portmap daemon running?

Is /nfsroot/eon exported in /etc/exports?
Author:  garetjax67 [ Wed Sep 05, 2007 9:04 pm ]
Post subject: 
Thanks Kirk. I looked at netstat and the box is listening on 111, but I went ahead and restarted portmap.

The following line is in my /etc/exports
Code:
/nfsroot/eon *(rw,no_root_squash,async)


Unfortunately, no change on the client. Any other thoughts? I really appreciate all of the help so far.
Author:  Kirk [ Wed Sep 05, 2007 9:43 pm ]
Post subject: 
Ok, that looks fine.. you can confirm that it is actually exporting properly using another box, or a live CD or something.

I'm thinking it's still about the IP stuff, cause the client actually gets the IP twice during boot, once when it connects to the DHCP server, and once again when it loads up the kernel.. and I think it's dropping it here.

Can you show us the /nfsroot/pxelinux.cfg/default file and/or 00-e0-c5-59-61-c3 file please?

We'll figure it out soon enough!
Author:  garetjax67 [ Sat Sep 08, 2007 11:22 am ]
Post subject: 
Hey all sorry for the delay -- life interrupted for awhile.

I tried a Dell GX150 and that sort of worked as a client so I am guessing that the export is working.

here is the contents of the /tftpboot/pxelinux.cfg/default file

Code:
DEFAULT vmlinuz-nfsroot
APPEND root=/dev/nfs nodhcp nfsroot=192.168.11.11:/nfsroot/eon


Thanks.[/code]
Author:  Kirk [ Sun Sep 09, 2007 7:04 pm ]
Post subject: 
Gee, I'm starting to run out of ideas, heh.

Let's try explicitly setting the IP and all the other parameters so that it shouldn't need any DHCP. Change your default file to;

Code:
DEFAULT vmlinuz-nfsroot
APPEND root=/dev/nfs ip=192.168.11.18:192.168.11.11:192.168.11.1:255.255.255.0::: nfsroot=192.168.11.11:/nfsroot/eon

IP notation is the following;
Code:
ip=<client-ip>:<server-ip>:<gateway-ip>:<netmask>:::

The three colons afterwards are important.

Fingers crossed!
Author:  garetjax67 [ Sun Sep 09, 2007 9:06 pm ]
Post subject: 
Kirk thanks for all your help, it did not appear to work but it did change the output on the client
Code:

NETDEV WATCHDOG: eth0: transmit timed out
eth0: Transmit timed out, status 0003, PHY status 786d, resetting. . .
ehto: link up, 100Mbps, full-duplex, lpa 0xCDE1
mount: server 192.168.11.11 not responding, timed out
Root-NFS: server returned error -5 while mounting /nfsroot/eon
VFS: Unable to mount root fs via NFS, trying floppy.
VFS: Insert root floppy and press ENTER



Does this spark any ideas?
Author:  Kirk [ Sun Sep 09, 2007 10:29 pm ]
Post subject: 
I plugged "eth0: Transmit timed out, status 0003, PHY status 786d, resetting. . . " into google and found a couple of references of it leading to a possible IRQ problem.

A couple of people tried adding "noapic" or "acpi=noirq" or "pci=noacpi" to the APPEND line and it worked with various successes. Another said he disabled SATA in the BIOS and it worked.

http://www.linuxquestions.org/questions/showthread.php?t=492159

Try pumping a couple of the other errors into google as well and see if you can come up with anything more concrete. It could be a kernel issue, or perhaps a driver issue?

Dunno if you could fit a different network card in those things?

Good luck!
Page 1 of 1 All times are UTC - 6 hours
Powered by phpBB® Forum Software © phpBB Group
http://www.phpbb.com/