|
osdir.com mailing list archive |
|
Subject: Re: Netbooting - Restricting Nic probing - msg#00010List: boot-loaders.grub.bugs> Guys - Have trawled through the archives but can't track down any > similar issues - atleast reported against recent versions of grub. > I'll try and be succinct. > I'm netbooting Dell 1650's/2650's/2850's using dhcp/tftp/pxegrub - > pxegrub is compiled with e1000 driver. > Each server has eth1 wired to the network that is served by DHCP/TFTP. > Eth0 is served by a seperate subnet. > > At boot time eth1 is succesfully assigned an IP and tftp downloads > pxegrub no problem. However, once loaded pxegrub probes and finds > eth0 (and any other nic that can be initialized by the driver). This > seems to inhibit pxegrub to download menu.lst to the correct nic of > eth1. Subsequently the boot process hangs. > > Is there a way to restrict the probing? No. All you can do is tweak the source code. I proposed a solution for this some years ago, but I didn't implement it, because I got no response, so I thought there was no interest. This should be addressed in GRUB 2. Okuji Thread at a glance:
Previous Message by Date: click to view message previewRe: Is there a nofb equivalent for grub menu display?On Monday 02 May 2005 23:15, Reshat Sabiq wrote: > My question is about the grub splash screen. When it's enabled in this > setup, no menu, in fact nothing, can be seen during boot, although if you > ht space, and then Enter it is equivalent to going to the menu and booting > the default option. I wonder if setting something like a nofb flag in grub > would make the splash screen work, but i suspect there is no such flag. We do not have a splash screen. Ask your vendor. Okuji Next Message by Date: click to view message previewRe: [PATCH] GRUB, XFS and writable strings == segfaultOn Friday 29 April 2005 01:57, Sven Wegener wrote: > First we declare usual to contain two constant strings. Then we assign > one of them to name and later we (might) try to write to one character > of the string. Please find attached a patch against current CVS that > transforms usual into an array of characters, so that writing to them is > safe. Thank you for your report. I have checked in a different fix for this. Please do not forget to write a ChangeLog entry next time. Okuji Previous Message by Thread: click to view message previewVFS: Unable to mount root fs on unknown-block(0,0)Problem: I created my own custom Linux kernel (2.6.11-8). I've compiled pretty much everything that has no reason to leave this box into the kernel (SATA drivers, fs [ext2 and ext3], etc). To compile the kernel I simply, > cd /usr/src/linux > make menuconfig (then do my configing) > make clean bzImage modules > make modules_install > make install make install creates the initrd for me (which always eeems to be one of the "solutions" to people's problems.. apparently not for me) and even modifies my grub.conf using grubby. Yet, when I boot, I continue to get the output listed in the topic during startup. At the stage in the game, the kernel is mounting all the necessary partitions it needs and grub should long be gone. I have read through the GNU Grub online manual (twice now.. though I'm still not really sure how much works), many google searches of proposed solutions to this problem, looked through RedHat's source rpm spec file %post tag, and I've still been unable to figure out why this is occurring. Since editing my /etc/grub.conf and replacing 'root=LABEL=/' with 'root=/dev/sda7' WORKS.. I'm trying to figure out why using LABEL=/ does not. Here's some discovery info I'm providing in case it's needed. At the grub prompt during boot, pressing 'c': grub> root (hd0,0): Filesystem type is ext2fs, partition type 0x83 =============== The entry I'm attempting to use: root (hd0,0) kernel /vmlinuz-2.6.11.8-custom ro root=LABEL=/ initrf /initrd-2.6.11.8-custom.img =============== Using the pre-installed kernel that came with this Fedora Core 2 updated release, then at the console typing, grub --no-floppy (this box doesn't have a floppy): grub> root (fd0): Filesystem type unknown, partition type 0x0 grub.conf: =============== > cat /boot/grub/grub.conf #boot=/dev/sda default=0 timeout=10 splashimage=(hd0,0)/grub/splash.xpm.gz title Fedora Core (2.6.11.8-custom) root (hd0,0) kernel /vmlinuz-2.6.11.8-custom ro root=LABEL=/ initrd /initrd-2.6.11.8-custom.img title Fedora Core (2.6.5-1.358smp) root (hd0,0) kernel /vmlinuz-2.6.5-1.358smp ro root=LABEL=/ initrd /initrd-2.6.5-1.358smp.img title Fedora Core (2.6.5-1.358) root (hd0,0) kernel /vmlinuz-2.6.5-1.358 ro root=LABEL=/ initrd /initrd-2.6.5-1.358.img =============== > cat /boot/grub/device.map # this device map was generated by anaconda (hd0) /dev/sda =============== > cat /etc/fstab LABEL=/ / ext3 defaults 1 1 LABEL=/boot /boot ext3 defaults 1 2 etc... I've been working on this out for about 1.5 days now. Could anyone lend a helpful hand? Regards, Kurt Next Message by Thread: click to view message previewRe: [PATCH] GRUB, XFS and writable strings == segfaultOn Friday 29 April 2005 01:57, Sven Wegener wrote: > First we declare usual to contain two constant strings. Then we assign > one of them to name and later we (might) try to write to one character > of the string. Please find attached a patch against current CVS that > transforms usual into an array of characters, so that writing to them is > safe. Thank you for your report. I have checked in a different fix for this. Please do not forget to write a ChangeLog entry next time. Okuji Loading Comments...
|
|
