Hey guys, had to chop my response into two parts as the message was too big:
Charles, I looked into your suggestion, but upon examining the tcpdump through ethereal, I can see that a DHCP discover is being sent (with accompanying appropriate MAC address for eth1 in the Ethernet II section of its description.) The only thing that I can think that strikes me as strange in the dump are two things. One is that the packet that sends the DHCP Discovery out is listed (in ethereal) as having a source of 0.0.0.0. This didn't concern me as a similar packet that handles the initial DHCP Discover (to start up the network boot in the first place) has the same source listed but works (this is, of course before the initrd and kernel are loaded.) The second thing that seemed interesting was that right before the DHCP Discovery attempt was made (from the loaded systemimager initrd and kernel) there was a Cisco Dis covery Protocol (of CDP protocol), then a Dynamic Trunking Protocol (of DTP protocol), then two replies (of LOOP protocol). Not sure if this would be affecting anything or if its part of normal operation of the two cisco switches we have on the chassis.
Ezra, I had tried to get the grub-installer to work after I had pushed the image in its entirety (with my hacked version of initrd.img, kernel, and boel binaries) but there were some weird messages being displayed, basically it didn't recognize that hd0 was a valid device when I tried to run it (I also tried the grub prompt and it didn't recognize hd0 as a valid drive when I ran the root (hd0,0) or setup (hd0) commands.) I'm assuming this has something to do with how the initrd from systemimager arranges the filessystem and how grub won't probe the bios devices correctly (I read some posts about that, but the consensus was that grub doesn't always probe correctly.) Is there a way around this? Like I said, I'm able to push the whole image to the blade with 1.06, it's just when I do the reboot after push it hangs at the GRUB message. I can get to a prompt by hacking the systemimager .master script by inserting a sleep command (then exiting to the prompt with ctrl-c.) When I go in and look at the machine, I see the entire image pushed properly (all the files seem there) under the /a/ directory. But, like I had posted, attempts at using grub and grub-install result in the hd0 not being detected (which I think goes back to the grub not probing correctly.) Is there a way around this? A way I can get grub or grub-install to detect the hd0?
Thanks for the suggestions and help, I appreciate it.
Steve
Charles, I looked into your suggestion, but upon examining the tcpdump through ethereal, I can see that a DHCP discover is being sent (with accompanying appropriate MAC address for eth1 in the Ethernet II section of its description.) The only thing that I can think that strikes me as strange in the dump are two things. One is that the packet that sends the DHCP Discovery out is listed (in ethereal) as having a source of 0.0.0.0. This didn't concern me as a similar packet that handles the initial DHCP Discover (to start up the network boot in the first place) has the same source listed but works (this is, of course before the initrd and kernel are loaded.) The second thing that seemed interesting was that right before the DHCP Discovery attempt was made (from the loaded systemimager initrd and kernel) there was a Cisco Dis covery Protocol (of CDP protocol), then a Dynamic Trunking Protocol (of DTP protocol), then two replies (of LOOP protocol). Not sure if this would be affecting anything or if its part of normal operation of the two cisco switches we have on the chassis.
Ezra, I had tried to get the grub-installer to work after I had pushed the image in its entirety (with my hacked version of initrd.img, kernel, and boel binaries) but there were some weird messages being displayed, basically it didn't recognize that hd0 was a valid device when I tried to run it (I also tried the grub prompt and it didn't recognize hd0 as a valid drive when I ran the root (hd0,0) or setup (hd0) commands.) I'm assuming this has something to do with how the initrd from systemimager arranges the filessystem and how grub won't probe the bios devices correctly (I read some posts about that, but the consensus was that grub doesn't always probe correctly.) Is there a way around this? Like I said, I'm able to push the whole image to the blade with 1.06, it's just when I do the reboot after push it hangs at the GRUB message. I can get to a prompt by hacking the systemimager .master script by inserting a sleep command (then exiting to the prompt with ctrl-c.) When I go in and look at the machine, I see the entire image pushed properly (all the files seem there) under the /a/ directory. But, like I had posted, attempts at using grub and grub-install result in the hd0 not being detected (which I think goes back to the grub not probing correctly.) Is there a way around this? A way I can get grub or grub-install to detect the hd0?
Thanks for the suggestions and help, I appreciate it.
Steve
Yahoo! Photos
Ring in the New Year with Photo Calendars. Add photos, events, holidays, whatever.
