Re: ESX4i and gPXE boot
Stefan Hajnoczi <[email protected]> Tue, 15 Dec 2009 09:22:34 +0000
| Newsgroups | gmane.network.etherboot.user |
|---|---|
| Message-ID | <[email protected]> |
On Mon, Dec 14, 2009 at 4:03 PM, Viswanath Krishnamurthy <[email protected]> wrote: >>> 2. When optionROM is burnt to a card, gPXE hangs when trying to load >>> image.tgz file (which is about 170MB). It consistently fails here and >>> older >>> versions seemed to work fine. (COMBOOT and Multiboot is disabled in >>> gPXE). >>> mbbot.c32 is the one shipped in ESX4i. I am wondering if gPXE memory is >>> getting overwritten by the image being downloaded causing this behavior. You know the old version works, so we can identify the code change that introduced the bug using git-bisect(1). Bisection involves testing gPXE versions between the known good and bad points to find the first version where the hang occurs. The bisection process is efficient because it tests the mid-point between the known good and bad points rather than forcing you to test every single version, one at a time. It shouldn't take long to do this and the result will be the git commit which broke your use case. The full man page for git-bisect(1) is here: http://www.kernel.org/pub/software/scm/git/docs/git-bisect.html There are tutorials on git-bisect(1) you may be interested in: http://jbowes.wordpress.com/2007/02/18/git-bisect-a-practical-example-with-yum/ http://www.reactivated.net/weblog/archives/2006/01/using-git-bisect-to-find-buggy-kernel-patches/ http://www.winehq.org/docs/winedev-guide/x1348 You will need git and working build environment. You can grab the gPXE git tree like this: git clone 'git://git.etherboot.org/scm/gpxe.git' If you would help doing the git-bisect(1), please join the #etherboot IRC channel on chat.freenode.net. Stefan ------------------------------------------------------------------------------ Return on Information: Google Enterprise Search pays you back Get the facts. http://p.sf.net/sfu/google-dev2dev