[gPXE] [Etherboot-discuss] D-Link DGE-530T 10/100/1000

[Shao Miller]

I've carbon-copied the gPXE mailing list, since this is gPXE-related. 
http://etherboot.org/mailman/listinfo/gpxe

fab wrote:
>
> net0: 00:1e:58:FF:FF:FF on PCI00:12.0 (open)
> [link:down, TX:0 TXE:0 RX:0 RXE:0
> link status: Not connected (0x38086001)]
>
> Waiting for link up on net0... ok
> DHCP (net0: 00:1e:58:FF:FF:FF)........
> connection timed out 0x4c106035)
> No more network device
>
There are a few ways in which this can go wrong.

If the DHCP service does not like gPXE, then you might not get an IP 
address.  An example of that would be the Nortel 470-48T layer 2 
switch.  If you think this might be the case, there is a version 
available for gPXE at 
http://git.etherboot.org/?p=gpxe-staging.git;a=shortlog;h=refs/heads/sha0-9-dhcp_relay_issue


Recently, I've had an issue on HP Compaq dc7600s (and relatives) where 
DHCP will work if initiated manually from the CLI, but not if run from 
an embedded script.  Enabling TIME_CMD in gpxe/src/config/general.h and 
using the 'sleep' command in the embedded script didn't seem to help.  
You could test to find out if you are having a similar issue to this by 
pressing Control-B at the Control-B prompt and then trying 'dhcp net0' 
and checking the results.

Ultimately, a tcpdump or Wireshark packet capture filtered for traffic 
from the gPXE client's MAC address and for the DHCP server (or its relay 
agent, if applicable) might reveal something that could complement the 
DEBUG= output that Thomas suggested.

- Shao Miller
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://etherboot.org/pipermail/gpxe/attachments/20091130/e867529c/attachment.html