I had the same problem at the beginning, but unfortunately I didn't keep
very detailed notes.
I don't think it's due to a single reason (may be not). Couple of things
to check: the line "+hardware.mod_eth0 alias eth0 [3c59x - replace with
your own]" in the node profiles is correctly typed so that it's not
defaulting to eepro100, which messed up all network access and the
SITE_NETWORK in site-cfg.h is correctly typed. I touched the 'fstab'
As logs on the nodes have all been overwritten, I'm not sure which one
fixed the problem. But somehow it was fixed.
Though it's fixed for me, one question to the group: This is from the
'fstab install /root' stage of the install (redhat73-cfg) where the cp
failed because /root//var/obj/ directory does not exist. At what point
are the directories in /root//* supposed to be populated?
Apologies if none of that is relevant .... If I remember anything more
specific I'll let you know.
On Fri, 2004-01-23 at 13:17, Steve Traylen wrote:
> Hi Mark,
> On Fri, 23 Jan 2004, Mark Nelson wrote:
> > Hi
> > I'm trying to install a CE from a configured lcfg, I've set up the lcfg
> > as layed out in the document on the GridPP website. When I boot the CE
> > node the install crashes out with the following error -
> > cp cannot create regular file `/root//var/obj/conf/fstab/fstab.hda': no
> > such file and directory
> > /etc/obj/fstab /root/etc/fstab.new: no such file and directory.
> > LCFG object fstab: Missing fstab template for hda
> > (/var/obj/conf/fstab/fstab.hda
> This is happening at Birminging currently, I have seen it before but
> forget now what caused it.
> When it fails at this point does it drop you into a shell?
> If so could your send back the output of both `mount` and
> /etc/obj/fstab adddisk hda /root
> > Investigation shows that the /root/etc directory is not there, I was
> > led to believe that the install procedure installs it's own custom
> > install of Redhat 7.3. It seems that the redhat install was not copied
> > down, there is what seems to be an install of Redhat sitting in
> > /opt/local/linux/nginstallroot/7.3/ on the lcfg server, I assume this is
> > copied down to the CE as part of the process. I've attached a copy of
> > the site-cfg.h file, I assume it is just me or is there a bug in the
> > install procedure, if there is something wrong with in install procedure
> > I'm willing to have a go at producing a fix for it.
> The OS located is /opt/local/linux/nginstallroot/7.3/ is mounted
> at install time as a root over NFS file system.
> LCFG in a nutshell mounts this as a root, formats the hard disk, mounts
> the hard disk on /root and calls rpm -Uvh *.rpm --root /root to actually
> get the OS from the original RPMS on to the disk.
> > I did do a /etc/init.d/mkxprofd reload prior to starting the install
> > Thanks in advance
> > Mark.
> > ----------------------------------------------------------------
> > Mark Nelson - [log in to unmask]
> > IPPP, Department of Physics, University of Durham,
> > Science Laborartories, South Road, Durham, DH1 3LE
> > Reception: +44 (0)191 334 3811, Direct Dial: +44 (0)191 334 3653
> > This mail is for the addressee only
> Steve Traylen
> [log in to unmask]