[PL #6755] updates - planetlab in a box discussion (bugs found)

[jaffe at cs.huji.ac.il via RT]

Email Recipients (see http://www.planet-lab.org/Support)
       Owner: alk
       Requestor: alk at absarokasoft.com
       Ticket Ccs: Thierry.Parmentelat at sophia.inria.fr, abenda at cs.huji.ac.il, alk at absarokasoft.com, alsbergt at cs.huji.ac.il, daniel51 at cs.huji.ac.il, devel-community at lists.planet-lab.org, jaffe at cs.huji.ac.il, jeff.sedayao at intel.com, nakao at iii.u-tokyo.ac.jp, renault.john at epfl.ch, smuir at cs.princeton.edu, winson.c.chan at intel.com

==================================================

Can someone explain about the "kexec bug".  Is this a matter of  
recompiling the kernel with a different set of flags, or is there a  
real bug in the 2.6.12 kernels?  On our latest bootcd build, we get  
all the way up to kexec and then we fail.

We NEED a 2.6.12 kernel on our bootcd's in order to fix a network  
driver bug, but this means that we can't bootstrap Planetlab, and  
hence we are still dead in the water.

Thanks,
Elliot

On Nov 28, 2005, at 4:26 PM, Mark Huang via RT wrote:

> Email Recipients (see http://www.planet-lab.org/Support)
>        Owner: alk
>        Requestor: alk at absarokasoft.com
>        Ticket Ccs: Thierry.Parmentelat at sophia.inria.fr,  
> abenda at cs.huji.ac.il, alk at absarokasoft.com, alsbergt at cs.huji.ac.il,  
> daniel51 at cs.huji.ac.il, devel-community at lists.planet-lab.org,  
> jaffe at cs.huji.ac.il, jeff.sedayao at intel.com, nakao at iii.u- 
> tokyo.ac.jp, renault.john at epfl.ch, smuir at cs.princeton.edu,  
> winson.c.chan at intel.com
>
> ==================================================
>
> Thierry Parmentelat via RT wrote:
>> The current code still needs changes.
>> - at the very least build.sh and update_rpm_repo.sh need to pull/ 
>> update
>> the new planetlab/ repo
>
> Even when the rollout is complete, the 3.2 kernel will not work for  
> the
> boot CD because of the kexec bug. We'll need to figure out what we  
> want
> to do to get a new boot CD pushed out, probably we'll want to  
> branch again.