[prev in list] [next in list] [prev in thread] [next in thread] 

List:       npaci-rocks-discussion
Subject:    Re: [Rocks-Discuss] Rocks 5.1 node reinstall pxe hangs at "select a
From:       Dussurget Frederic <Frederic.Dussurget () ish-lyon ! cnrs ! fr>
Date:       2010-06-30 14:50:21
Message-ID: 4C2B59AD.2030904 () ish-lyon ! cnrs ! fr
[Download RAW message or body]

Hi there,
My Cluster is online again !
Thanks to Zheng Li, I had to chmod o+r and o+x the /root folder on the 
frontend. (I guess the kickstart script needs access to the /root/.ssh 
files ... ?)
I wonder if it is safe to keep those permissions on /root ?
Anyway, thanks a lot :-)
Fred

Le 29/06/2010 18:36, Greg Bruno a écrit :
> On Tue, Jun 29, 2010 at 8:31 AM, Dussurget Frederic
> <Frederic.Dussurget@ish-lyon.cnrs.fr>  wrote:
> 
> > Most urgent thing, I send a tentakel command before checking if nodes
> > could reinstall: (I know I was really stupid)
> > 
> > /tentakel -g compute '/boot/kickstart/cluster-kickstart-pxe'
> > 
> > /Is there a way to cancel this tentakel command, because all the nodes rebooted \
> > and wants to reinstall ! There's zero node on-line and avalaible now ... I'm not \
> > really happy with the idea of upgrading rocks now because I'm running short of \
> > time these days and the cluster has to be online asap :-((( I'd like to upgrade \
> > in september if possible ... 
> > 
> connect a monitor and keyboard up to one of the compute nodes. at the
> grub boot screen, press a key to stop the boot. then, you should be
> able to hit the arrow down key and select a 'normal boot' (rather than
> a reinstallation boot).
> 
> as for why you can kickstart a node, first, see if you can
> successfully create a kickstart file:
> 
> # rocks list host profile compute-0-0>  /tmp/ks.cfg
> 
> - gb
> 
> 
> 


[prev in list] [next in list] [prev in thread] [next in thread]